Datasets:

shreyaspimpalgaonkar

/

ycombinator_s24

like 0

Silurian

silurian

https://silurian.ai/

Kirkland, WA, USA

Silurian is building foundation models for simulating Earth, starting with weather. From assessing the risk of wildfires to predicting the energy grid load, we provide an infrastructure layer for our planet. Our frontier models push the boundaries of what can be simulated on Earth and improve decision making across vital sectors including energy, insurance, agriculture, and logistics.

Foundation models to simulate Earth

Industrials

Industrials -> Climate

S24

Active

Early

https://www.ycombinator.com/companies/silurian

https://yc-oss.github.io/api/batches/s24/silurian.json

{"code":"ETIMEDOUT","name":"Error","message":"Timeout waiting for dependencies(PuppeteerControl) to be ready for CrawlerHost."}

Kisho

kisho

https://kisho.app

San Francisco, CA, USA

Kisho is an AI data scientist that lives inside a Jupyter Notebook. It enables anyone to perform advanced data analysis and build ML models using natural language - no coding required.

The AI Data Scientist - a Jupyter Notebook that writes itself.

B2B

Unspecified

S24

Active

Early

https://www.ycombinator.com/companies/kisho

https://yc-oss.github.io/api/batches/s24/kisho.json

Title: Y Combinator | File Not Found URL Source: Warning: Target URL returned error 404: Not Found Markdown Content: 404 --- [Back to the homepage]( support please contact [[email protected]](mailto:[email protected])

MinusX

minusx

https://minusx.ai

San Francisco, CA, USA; Remote

MinusX is a chrome extension that adds a side chat to your analytics apps (Jupyter, Metabase, Grafana, Tableau, etc). Given an instruction, our agent operates your apps - by clicking & typing, just like you do - to analyze data and answer queries. We believe an AI Data Scientist is a scientist, not yet-another-new-analytics-platform. MinusX interoperates with you in tools you already love and use, and as a matter of philosophy, gets out of the way.

AI Data Scientist for Jupyter and Metabase

B2B

B2B -> Analytics

S24

Active

Early

https://www.ycombinator.com/companies/minusx

https://yc-oss.github.io/api/batches/s24/minusx.json

Title: MinusX: AI Data Scientist for Jupyter and Metabase | Y Combinator URL Source: Markdown Content: ### AI Data Scientist for Jupyter and Metabase MinusX is a chrome extension that adds a side chat to your analytics apps (Jupyter, Metabase, Grafana, Tableau, etc). Given an instruction, our agent operates your apps - by clicking & typing, just like you do - to analyze data and answer queries. We believe an AI Data Scientist is a scientist, not yet-another-new-analytics-platform. MinusX interoperates with you in tools you already love and use, and as a matter of philosophy, gets out of the way. MinusX Founded:2024 Team Size:3 Location:San Francisco ### Active Founders ### Vivek Aithal, Founder Co-founder & CEO of MinusX. Spent the last 3 years as a research engineer building the world's best open-source autonomous driving agent at comma.ai. Currently building a data science agent that can sit in your browser and use all your data tools. UC Berkeley and IIT Kharagpur grad. ### Sreejith Puthanpurayil, Founder Co-Founder & CTO of MinusX. Previous stints include Ads Delivery & Ranking at Meta, Ride-hailing backend & infra at Gojek, & full stack at Elanic, a social c2c marketplace (acquired by Share Chat) ### Arpit Saxena, Founder Ex-Google (Google Pay), GreyOrange Robotics, Udaan.com; Software and Systems Engineer building warehouse automation solutions for km sized warehouses ### Company Launches [### MinusX: AI Data Scientist for Jupyter and Metabase]( **TL;DR** --------- MinusX is a Chrome Extension that adds a side chat to Jupyter and Metabase. Given an instruction, our agent operates the apps— via clicking and typing— to analyze data and answer queries. We strongly believe that you already have all the tools you need; you just need someone like MinusX to use them! Checkout [**minusx.ai**]( for more info. [ **Asks** -------- 1. Install our [**Chrome extension**]( 2. Take MinusX for a spin in our [**playground**]( 3. We’re working on supporting more tools. Want us to hurry up, or don't see your favorite tool in our list? Here’s a [**google form**]( you can fill out so that we can notify you when we support your tool! **Features** ------------ ### 1\. Generate Hypotheses and Explore data ### 2\. Interop with MinusX to modify existing Jupyter notebooks or Metabase Questions ### 3\. Select a region and ask questions, or ask for modifications **Pain** -------- * Lack of data analyst/data scientist bandwidth is a certified pain. * If you’re a programmer, you just want answers. If you’re a product manager, you just want answers. If you’re an analyst/scientist, you want 10 clones of yourself. * Any new fancy “talk-to-data” platform is a whole thing. You have to migrate your data, and you have to convince your whole team to move— just to start using it. * Working in ChatGPT / Claude requires constant copy-pasting. Also, you want to inter-operate with AI, make corrections when it messes up, and guide it rather than just admonish it in chat. * More importantly, they’re still just chat! No AI system today can really use analytics software - click and type - to get you that data you needed 5 minutes ago. **Our Approach** ---------------- > ### **_An AI Data Scientist is a Scientist, not yet-another-new-analytics-platform_** * Advanced agents should just work with tools you already use and get out of the way. * MinusX integrates seamlessly into workflows. You can invoke it only when you need it to do something. * Whether you’re a data analyst, programmer, or product manager, you can use MinusX to speed up your workflows instantly. * We believe the path to advanced agents runs through specialized, useful intermediaries. **Team** -------- We love to build stuff. We met about a decade ago in the drowsy hinterlands of IIT Kharagpur and have been building things together at work, on side projects, and at hackathons ever since. [**Vivek**]( was most recently a Research Engineer working on self-driving cars at [**comma.ai**]( [**Sreejith**]( was a Senior Engineer in the Meta shops ads ranking team optimizing Facebook Shops Ads delivery & core infrastructure, and [**Arpit**]( was building a warehouse robot fleet as a Product Engineer at Udaan. We have extensive experience in machine learning, data infrastructure, and systems engineering needed to build the best, most delightful class of Data Science agents you'll ever use. **Ask(ing again!)** ------------------- 1. Install our [**Chrome extension**]( 2. Take MinusX for a spin in our [**playground**]( 3. We’re working on supporting more tools. Want us to hurry up, or don't see your favorite tool in our list? Here’s a [**google form**]( you can fill out so we can notify you when we support your tool! If you have any questions or comments, you can join our [**discord**]( or ping me anytime at [**[email protected]**](mailto:[email protected]). Hope you enjoy using MinusX as much as we enjoy building it! ### ❤️

Pharos

pharos

https://pharos.health/

San Francisco, CA, USA

Pharos automates hospital quality reporting, saving millions in labour costs and helping to prevent avoidable patient harm. Today, clinicians spend thousands of hours manually pulling complex facts out of medical records for mandatory reporting and quality improvement. Our AI pulls those facts out of unstructured medical records automatically. We automate reporting and show staff where avoidable patient harm is happening.

Automate hospital reporting and prevent patient harm with AI

Healthcare

Healthcare -> Healthcare IT

S24

Active

Early

https://www.ycombinator.com/companies/pharos

https://yc-oss.github.io/api/batches/s24/pharos.json

Title: Pharos: Automate hospital reporting and prevent patient harm with AI | Y Combinator URL Source: Markdown Content: ### Automate hospital reporting and prevent patient harm with AI Pharos automates hospital quality reporting, saving millions in labour costs and helping to prevent avoidable patient harm. Today, clinicians spend thousands of hours manually pulling complex facts out of medical records for mandatory reporting and quality improvement. Our AI pulls those facts out of unstructured medical records automatically. We automate reporting and show staff where avoidable patient harm is happening. Pharos Founded:2024 Team Size:3 Location:San Francisco ### Active Founders ### Felix Brann, Founder CEO at Pharos. Previously VP Data Science at vital.io and VP Quantitative Research at JP. Morgan. Obsessed with sepsis: ### Matthew Jones, Founder Founder and CTO at Pharos. Previously I was part of the founding team of Market2x, a rural trucking SAAS startup, growing that from inception to international expansion. The rest of my career has been as a software engineer at various health tech companies. ### Company Launches [### 🏥 Pharos - Automate reporting and prevent patient harm with AI]( **tl;dr:** _The data hospital teams need to improve patient safety is buried in unstructured medical records. Today, clinicians spend thousands of hours manually ‘abstracting’ it for reporting and analysis. We automate the entire process and use the data to show them where and why avoidable harm is happening._ Hi folks! We’re Felix and Matthew, and we’re building [Pharos]( The problem: ------------ Avoidable harm happens in hospitals **all the time**. Wards are busy, clinician turnover is high, and an aging population means increasingly complex patients. Sepsis alone kills 350,000 patients a year in the US, and a significant number of those deaths are preventable. Hospitals have teams dedicated to preventing harm. They track avoidable events, identify the process failures that cause them, and report performance data to clinical registries. This means identifying harm events, risk factors and process adherence from patient journeys composed of pages of unstructured clinical notes. Today, this is an entirely manual process. Producing structured quality metrics from a single complex patient case can take up to **8 hours of clinical time**. A single hospital can spend **$5m per year** extracting this data, and it still arrives weeks after discharge, on a small sample of their patients. The solution: ------------- Our AI extracts the data quality teams need from **every patient record in real-time**. It produces verifiable quality metrics, with references into the original medical record. We use this data to: * **Automate reporting** for clinical registries and value-based reimbursement contracts, saving thousands of clinical hours. * **Identify and surface process failures** that are contributing to patient harm, letting teams take action on issues like sepsis, hospital-acquired infections, and pressure ulcers. * **Measure the impact of quality improvement projects** in real-time rather than months after implementation. Why us? ------- [Felix]( and [Matthew]( spent the past 5 years deploying patient and clinician-facing AI into over 70 hospitals together. As VP of Data Science, Felix published papers in major medical journals on sepsis prediction and medical record summarization using LLMs. Matthew has years of experience integrating software into EHRs and previously built another startup from inception to international expansion. Alex joined the team after working as a doctor in the UK and then as a medical AI researcher at Imperial College London and Meta’s Reality Labs. He experienced this problem firsthand, spending years of his residency frustrated at the manual abstraction required for quality improvement. We believe enabling quality teams with AI represents a huge opportunity to save lives and prevent harm. Our ask: -------- Please reach out to [[email protected]](mailto:[email protected]) if you know the following people! * Anyone working at a senior level at a US hospital (we’ll ask them for an intro to their quality team) * Anyone working in healthcare with a title that includes “Quality”, “Patient Safety,” or “(Sepsis, Stroke, …) Coordinator” * Academics and clinicians working at the intersection of data and clinical quality ### Hear from the founders #### How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?) Felix and Matthew worked together for nearly 5 years at [Vital.io]( a company deploying AI models into hospitals. While piloting clinical adoption of a predictive sepsis model, they realized that enabling quality staff with AI represents a huge opportunity to improve patient outcomes. #### What is your long-term vision? If you truly succeed, what will be different about the world? We believe in a future where AI catches medical mistakes everywhere in the hospital, before they become serious. We want to be the lighthouse for our hospitals, supporting clinicians and reducing patient harm.

Distro

distro

https://distro.app

New York, NY, USA

Distro is the all-in-one AI-powered sales platform for counter staff and inside sales at industrial wholesale distributors (HVAC/R, plumbing, electrical, and more).

The AI co-pilot for sales reps at industrial wholesale distributors.

B2B

B2B -> Supply Chain and Logistics

S24

Active

Early

https://www.ycombinator.com/companies/distro

https://yc-oss.github.io/api/batches/s24/distro.json

Title: Distro: The AI co-pilot for sales reps at industrial wholesale distributors. | Y Combinator URL Source: Markdown Content: Distro: The AI co-pilot for sales reps at industrial wholesale distributors. | Y Combinator =============== []( "Y Combinator") [About]( [What Happens at YC?]( Interview Guide]( Blog]( [Companies]( [Startup Directory]( Companies]( Directory]( YC]( [Startup Jobs]( [All Jobs]( Engineering]( Operations]( Marketing]( Sales]( Internship Program 2024]( Job Guide]( Coaching]( Startup Jobs Blog]( [Find a Co-Founder]( [Library]( [SAFE]( [Resources]( [Startup School]( Investors]( News]( Open main menu Apply for **W2025** batch.[Apply]( "Apply for W2025 batch.") [Home]( Distro ====== The AI co-pilot for sales reps at industrial wholesale distributors. [S24]( Active [artificial-intelligence]( York]( * * * [Company]( [Jobs](  [ * * * ### The AI co-pilot for sales reps at industrial wholesale distributors. Distro is the all-in-one AI-powered sales platform for counter staff and inside sales at industrial wholesale distributors (HVAC/R, plumbing, electrical, and more). Distro Founded:2022 Team Size:4 Location:New York Group Partner:[Nicolas Dessaigne](  []( "LinkedIn profile") ### Active Founders ### Jason Sullivan, Founder Jason is the founder and CEO of Distro, the all-in-one AI co-pilot for counter staff and inside sales at industrial wholesale distributors. He is a repeat founder, having previously co-founded Vested. Before Vested, he was CTO at Elementus, Tech Lead at Coatue, and Senior Software Engineer at PeerIQ. In a prior life, Jason was a credit derivatives trader. Jason holds an MS in Computer Science from Stanford and a BA in Cognitive Science from Yale. Jason Sullivan [Distro](  []( "LinkedIn profile") ### Company Launches [### 🏭🤖 Distro - Bringing point-of-sale AI to industrial distribution]( [Distro]( is the AI sales co-pilot designed for industrial wholesale distributors. We created Distro out of a passion for developing software for the “real economy,” and in particular for the trades. The Distro team is based in, and loves, NYC! 🗽 ### 😅 **The Problem** Distributors, particularly those mid-size and smaller, face a long list of point-of-sale challenges these days. _🧠 Brain drain_: the most experienced reps are at or near retirement and are taking a great deal of tribal knowledge with them. _🔧 Complex sales_: quoting entire systems, load calculations, replacing legacy parts and equipment. _📦 Inventory challenges_: growing SKU counts, supply chain disruptions, decreased inventorying by contractors. _🎓 Training and retention_: counter reps, in particular, are difficult to hire, train and retain. These issues are amplified by the increasing importance of counter and inside sales relative to the rest of the sales organization–there has been a whopping 13 percentage point shift\* towards counter staff and inside sales influencing purchasing decisions, moving away from territory managers/outside sales. To top it all off, there is the ever-present and growing threat of larger players that have the resources to throw money and bodies at the aforementioned problems. _\*_ 2023-2024 State of the Channel, HARDI (Heating, Air-conditioning & Refrigeration Distributors International) ### 🔍 **Our Approach** What began as an academic curiosity about how to apply AI to distribution evolved into our team spending countless hours with distributor reps across the country over the course of a year, going deep on their day-to-day challenges, and building a distributor-centric solution from the ground up in partnership with several leading distributors. This experience–fully immersing ourselves in a distribution vertical in order to go far beyond a superficial understanding of that vertical’s product universe and sales workflows–has shaped our product philosophy and overall approach. Some of our learnings: 📍 _Meet reps where they are_: don’t try to push reps to adapt to a totally new paradigm. Adapt our workflows to how things really work at the counter. 🚀 _Effortless onboarding_: distributor IT teams don’t need another tech project to distract them. Build solutions that avoid the need for long implementations. 🔒 _Data sanctity_: distributors can spend a lot of time, energy, and money curating their data. Complete data security and privacy are necessary to protect that advantage. 🛠️ _Enhance, don’t replace_: the human element is the key to a distributor’s value-add to its contractor customers. We’re building Iron Man’s suit, not the Terminator. ### 💡 **Our Solution** Our platform helps counter and inside sales reps work more efficiently by transforming complex customer requests into real-time product information and quotes. With Distro, reps can quote faster, boost conversion rates, and increase customer retention. We support distributors in sectors like HVAC/R, plumbing, electrical, and more. ### 🙏 **Asks** 🤝 Connect with us: If you are a founder in the distribution, manufacturing, or contractor spaces, we’d love to share learnings. 📣 Customer intros: distributors, contractors, and OEMs, particularly in HVAC/R, plumbing, and electrical. Footer ------ [Y Combinator]( "Y Combinator") ### Programs * [YC Program]( * [Startup School]( * [Work at a Startup]( * [Co-Founder Matching]( ### Company * [YC Blog]( * [Contact]( * [Press]( * [People]( * [Careers]( * [Privacy Policy]( * [Notice at Collection]( * [Security]( * [Terms of Use]( ### Resources * [Startup Directory]( * [Startup Library]( * [Investors]( * [SAFE]( * [Hacker News]( * [Launch YC]( * [YC Deals]( ### Make something people want. [Apply]( [Twitter]( © 2024 Y Combinator

Stempad

stempad

https://www.stempad.io

You can think of Stempad as a Notion for science. It is the world's first true pen-and-paper alternative to fast scientific writing and collaborating. Quickly switch between different forms of technical visualization with the ease of a whiteboard and the convenience of your keyboard. Stempad allows you to share your work, collaborate in real time, store your data, annotate, write papers, plan, takes notes, create presentations, and so much more. Our vision is to make it easier and faster for students and scientists to digitize and share their scientific ideas.

Notion for science

B2B

B2B -> Productivity

S24

Active

Early

https://www.ycombinator.com/companies/stempad

https://yc-oss.github.io/api/batches/s24/stempad.json

Title: Stempad: Notion for science | Y Combinator URL Source: Markdown Content: ### Notion for science You can think of Stempad as a Notion for science. It is the world's first true pen-and-paper alternative to fast scientific writing and collaborating. Quickly switch between different forms of technical visualization with the ease of a whiteboard and the convenience of your keyboard. Stempad allows you to share your work, collaborate in real time, store your data, annotate, write papers, plan, takes notes, create presentations, and so much more. Our vision is to make it easier and faster for students and scientists to digitize and share their scientific ideas. ### Latest News Stempad Founded:2024 Team Size:1 Location: ### Active Founders ### Ralph Rouhana, Founder Hey, I'm Ralph 👋 I founded Stempad in early 2024 to help me take notes in my final year of school. After getting it to several hundred users and a few paying subscribers, I saw the potential for schools, research teams, pharmaceutical companies, tutoring companies, and more. I've previously interned at BioRender (W18), Microsoft, and 5 other companies. ### Company Launches [### Stempad: Scientific writing at the speed of thought]( **_Tl;dr:_** Stempad is an online editor and platform that streamlines writing and sharing scientific documents fast. Take notes, write research papers, create and conduct exams, collaborate in real-time, share your work with the world, and more. [_Try it out today._]( Hi everyone, I’m [Ralph]( I’m on a mission to improve the way teachers, students, and scientists document and share their ideas. **The Problem** --------------- Handwriting (via paper/tablet/board), a medium that is often substandard and that many struggle with, is currently the only viable way to do fast or impromptu scientific writing. * Students often resort to text editors unoptimized for science, such as Notion or Word, to take notes and write assignments. * Scientists use PowerPoint, wrestle with slow and expensive legacy software, and resort to handwriting to take notes, create writeups, or make presentations. * Remote tutors and their students require expensive tablets, styluses, and specialty software for a real-time digital whiteboard experience. * Professors often upload low quality scans of their messy handwritten class notes. The list could go on. The ability to quickly and collaboratively document scientific ideas with a keyboard is a huge QOL and productivity boost for people and institutions doing science. **Stempad To The Rescue 🔬** ---------------------------- You can think of Stempad as a **Notion for Science**. It is the world's first true pen-and-paper alternative for fast scientific writing. Quickly switch between different forms of technical visualization with the ease of a stylus and the convenience of your keyboard. Stempad allows users to save, export and share their work, collaborate in real time, create and grade assignments, conduct remote exams, write research papers, take notes, create presentations, and so much more. **My Asks** ----------- * Introduce me to educators, students, scientists, and decision makers **at schools, pharmaceutical companies, research teams, edtech companies, tutoring companies, summer schools, and relevant STEM programs.** [Click here](mailto:[email protected]?subject=Introducing%20you%20to%20Ralph%20at%20Stempad). * If you or any founders you know have B2B SaaS products in the education and science industry, I’d love to learn from you. [Click here](mailto:[email protected]?subject=Founder%20Intro). * [Sign u](mailto:[email protected]?subject=Founder%20Intro)[p]( use it next to others, and spread the word :)

Azalea Robotics Corporation

azalea-robotics-corporation

https://azalearobotics.com/

Berkeley, CA, USA

Azalea Robotics automates airport baggage handling with intelligent robot operations. The global market for airport baggage handling systems is $20+ billion and growing, presenting a significant opportunity for innovation and market disruption in this sector. Passenger air traffic volume is increasing, driving demand for efficient and reliable baggage handling at airports and putting immense pressure on existing infrastructure. In 2023 alone, airports processed approximately 4.5 billion bags, highlighting the need for advanced solutions to manage this load effectively. Azalea Robotics provides state-of-the-art robotic systems that enhance efficiency, reduce mishandling, and improve passenger experience through more reliable operations. Baggage handling is a critical component of airline ground operations, yet it is fraught with challenges. The work is physically demanding, often leading to long-term injuries among workers. Traditional baggage handling involves repetitive lifting and maneuvering of heavy loads, which can result in long-term health issues. Azalea Robotics addresses these challenges by automating the most strenuous tasks, thereby reducing the risk of injury and enhancing operational efficiency.

Automating airport baggage handling with robots.

Industrials

Industrials -> Manufacturing and Robotics

S24

Active

Early

https://www.ycombinator.com/companies/azalea-robotics-corporation

https://yc-oss.github.io/api/batches/s24/azalea-robotics-corporation.json

Title: Azalea Robotics Corporation: Automating airport baggage handling with robots. | Y Combinator URL Source: Markdown Content: ### Automating airport baggage handling with robots. Azalea Robotics automates airport baggage handling with intelligent robot operations. The global market for airport baggage handling systems is $20+ billion and growing, presenting a significant opportunity for innovation and market disruption in this sector. Passenger air traffic volume is increasing, driving demand for efficient and reliable baggage handling at airports and putting immense pressure on existing infrastructure. In 2023 alone, airports processed approximately 4.5 billion bags, highlighting the need for advanced solutions to manage this load effectively. Azalea Robotics provides state-of-the-art robotic systems that enhance efficiency, reduce mishandling, and improve passenger experience through more reliable operations. Baggage handling is a critical component of airline ground operations, yet it is fraught with challenges. The work is physically demanding, often leading to long-term injuries among workers. Traditional baggage handling involves repetitive lifting and maneuvering of heavy loads, which can result in long-term health issues. Azalea Robotics addresses these challenges by automating the most strenuous tasks, thereby reducing the risk of injury and enhancing operational efficiency. Azalea Robotics Corporation Founded:2023 Team Size:2 Location:Berkeley, CA ### Active Founders ### David Millard, Founder David is a roboticist and software engineer with a background in mathematics and computer science. He previously worked as a software engineer at Google X on the Everyday Robots project, Microsoft, and at IronOx (W16). David's doctoral research focused on robotic systems manipulating non-rigid objects and was funded by a NASA Space Technology Research Fellowship. He has published research with the NASA Jet Propulsion Lab, Ames Research Center, and Google DeepMind. ### John B. Stroud, Founder John B. is a finance and operations professional with experience in airline operations, management consulting, and large-scale operations finance. He has a full-time MBA from Kellogg School of Management at Northwestern University in finance and strategy. From his time at United Airlines, John B. understands the inner workings of major international airlines and the inherent challenges of running manual baggage handling systems that he brings to Azalea Robotics. ### Company Launches [### Azalea Robotics - Robots moving bags at airports]( **TL;DR** - [**Azalea Robotics**]( is building robots to handle baggage between touchpoints at airports. Our robots work around the clock, never get tired, and never lose or damage your bag after check-in. **The Team** - [David]( and [John B.]( met in undergrad at the University of Georgia and have been close friends for 12 years. David has a PhD in Computer Science from USC and is an expert in soft-body robotic manipulation. John B. has an MBA from Kellogg and extensive experience in airline ground operations. They founded Azalea Robotics at the intersection of their combined strengths to modernize bag-handling systems at airports. **The Problem** - Airport baggage handling is broken. It requires an immense amount of back-breaking labor, using decades-old infrastructure at airports, all while passengers increasingly choose to cram everything into a carry-on because they don’t trust airlines to deliver their luggage on time and in one piece. Airlines and airports have an increasingly hard time hiring these workers and are plagued by frequent system outages and customer confusion and complaints during large weather events. **The Solution** - Robotic systems work around the clock, never get tired and damage your bag, never misidentify bags, and are significantly cheaper to deploy than manual labor. Our technology allows large robotic arms to handle all bag types and intelligently improve with every bag transferred. Check out a sneak peek of our system in action! [ **Our Customers** - We sell our systems to airports and airlines around the world. Depending on the size of their airport presence (i.e., is this a hub for a given airline?), airlines will either outsource or hire in-house baggage handling for their customers. Airports are typically public entities run by the local city or metropolitan area and can be more or less hands-on with baggage handling, depending on their passenger volume. Worldwide, airports and airlines spend $20B+ to handle 4.5B+ bags (and growing) every year! **Our Ask** - Everyone has a nightmare baggage story, and we’d love to hear yours at [[email protected]](mailto:[email protected])! Otherwise, if you're interested in learning more or have someone you think we should talk to, let us know at [[email protected]](mailto:[email protected])!

Weavel

weavel

https://weavel.ai

San Francisco, CA, USA

Weavel automates prompt engineering, delivering the best prompts 50x faster than humans. Simply input your prompt and receive optimized prompts with highest accuracy. Boost your prompt's accuracy by an average 20% in less than 5 minutes. Andrew and Jun built 10+ LLM-based products, open-sourced a prompt engineering platform, and co-authored a paper at a NeurIPS workshop in 2023. Hyun Jie worked on data analytics and optimization at Chartmetric and DevRev, and focused on growth marketing at Liner.

Automate prompt engineering & get best prompts 50x faster

B2B

B2B -> Engineering, Product and Design

S24

Active

Early

https://www.ycombinator.com/companies/weavel

https://yc-oss.github.io/api/batches/s24/weavel.json

Title: Weavel: Automate prompt & LLM engineering, 50x times faster than a human | Y Combinator URL Source: Markdown Content: ### Automate prompt & LLM engineering, 50x times faster than a human Weavel automates prompt and LLM engineering, delivering best prompts and algorithms 50x times faster. We do this by using LLMs and search algorithms to replace the manual trial-and-error of prompt engineering, but more efficiently - it is 3 times faster than the leading open source project. Weavel Founded:2023 Team Size:3 Location:San Francisco ### Active Founders ### Andrew Chung, Founder Andrew is co-founder and CEO of Weavel. Andrew previously led product + engineering teams for two years, and has experience in building various applications in AI, VR, web, and mobile. Prior to that, Andrew studied electrical & computer engineering at Seoul National University before taking leave of absence in his junior year to focus on building things. ### Hyun Jie Jung, Founder HyunJie is co-founder of Weavel, specializing in growth and data analytics. She studied data science and media at UC Berkeley and has experience working at Liner, Chartmetric, and DevRev, where she focused on marketing and data analysis. ### Junyoung Park, Founder Jun is co-founder of Weavel. Jun previously worked as researcher at NLP-focused AI lab and has experience in building various applications with LLM. Prior to that, Jun studied computer science & engineering at Seoul National University. ### Company Launches [### Weavel - Ape 🐒 Your first AI Prompt Engineer]( **☕️ TL;DR** ------------ * **Ape** is the ultimate **AI prompt engineer** 🐒, designed to optimize your prompts by reducing cost and latency while increasing performance. * Ape achieves an **impressive 94.5%** on the GSM8K benchmark, surpassing Vanilla (54.5%), CoT (87.5%) and DSPy (90.0%). * **Easy to set up evaluation**: Ape can auto-generate evaluation code and use LLMs as a judge, or you can use your own eval metrics. * Get set up in less than 15 minutes and see the difference. * [Schedule a meeting]( to discover more. Let's chat! 🙂 **🔒 Problem** -------------- You’re an engineer of an LLM app, trying to get the prompts just right. Every time you type something in, the output changes—so you tweak a word here and there, and it changes again. Sometimes the outputs looks better, sometimes not. But you’re never sure. Hours go by, all spent on prompt engineering. Getting the outputs you want can feel like an endless game of trial and error. And you’re not alone. Over the past few weeks, we’ve talked to over 100 YC companies, and a lot of them are facing the same challenges: * **Measuring output quality is hard** (You’re heavily relying on manual evaluations at the moment.) * **Prompt engineering does not work as you want** (You hate spending 5-7 hours a day searching for that one great prompt.) **🔑 Solution** --------------- We solve the problem with one simple formula: ``` good input + right guidance = better prompts ``` Today, we launch Ape, your first **AI Prompt Engineer**. Inspired by DSPy, Reflexion, Expel and other research papers, Ape iteratively improves your prompts. Here’s how Ape works: 1️⃣ Log your inputs and outputs to Weavel (with a single line of code!) 2️⃣ Let Ape filter the logs into datasets. 3️⃣ Ape then generates evaluation code and uses LLMs as judges for complex tasks. 4️⃣ As more production data is added, Ape continues refining and improving prompt performance. ### **How to use** **Create a Dataset** Change just one line of code to start logging LLM calls with the Weavel Python SDK. The SDK supports sync/async OpenAI chat completions and OpenAI structured outputs. You can also import existing data or manually create a dataset. **Create a Prompt** Write a prompt that corresponds to your dataset. You can add an existing prompt as the base version, or if you prefer, create a blank prompt and provide a brief description for Ape to create a prompt from scratch. **Optimize Prompts** To optimize your prompt using Ape, fill in the necessary information (e.g. JSON schema as you want) and then run the optimization process. An enhanced version of your prompt will be created and available soon. Ta-da! It’s that easy. Ape outperforms with a remarkable 94.5% score on the GSM8K benchmark, surpassing Vanilla (54.5%), CoT (87.5%) and DSPy (90.0%). With Ape, you can optimize the prompt engineering process, saving tons of time and cost while increasing performance. Ape is **open source**. [Check out our repository on GitHub.]( (We’d appreciate a star 🌟) **🚀 The Team** --------------- From left to right: [**Jun**]( [**Andrew**]( [**HyunJie**]( and [**Toby**]( — together we’re building Weavel. **Andrew** and **Jun** built 10+ LLM-based products, open-sourced a prompt engineering platform, and co-authored a paper at a NeurIPS workshop last year. **HyunJie** worked on data analytics and optimization at Chartmetric and DevRev, and focused on growth marketing at Liner. Then **Toby** joined, a full-stack engineer who worked at several early stage teams, shipping 5+ products. **🙏 Ask** ---------- * Try Ape! [Schedule]( a walkthrough with the Weavel team or email [[email protected]](mailto:[email protected]). * Share thoughts on our [Discord]( or DM us on [Twitter]( * If you know anyone struggling with prompt engineering or evaluations for LLM apps, connect them with us! * Copy & paste blurb: A YC company named Weavel has developed an AI prompt engineer (Ape in short) which continuously improves your prompts. It’ll save tons of time for you. You can grab a time [here]( for a demo from the founders.

David AI

david-ai

https://www.withdavid.ai/

San Francisco, CA, USA

Data for multimodal AI

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/david-ai

https://yc-oss.github.io/api/batches/s24/david-ai.json

Title: David AI: Data for multimodal AI | Y Combinator URL Source: Markdown Content: ### Data for multimodal AI David AI Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Tomer Cohen, Founder Co-founder and CEO at David AI. Previously, Chief of Staff at Scale AI and Consultant at McKinsey & Company. ### Ben Wiley, Founder Co-founder and CTO at David AI. Previously, at Scale AI - Head of Engineering for Scale’s Public Sector GenAI Platform. Before that - SWE at Microsoft.

Zephr

zephr

https://www.zephr.ai/

San Francisco, CA, USA

Zephr is an AI customer success manager that lets you manage hundreds of customers with ease. Our AI monitors customer health, identifies expansion opportunities and churn, and automates repetitive tasks.

AI customer success manager

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/zephr

https://yc-oss.github.io/api/batches/s24/zephr.json

Title: Zephr: AI customer success manager | Y Combinator URL Source: Markdown Content: ### AI customer success manager Zephr is an AI customer success manager that lets you manage hundreds of customers with ease. Our AI monitors customer health, identifies expansion opportunities and churn, and automates repetitive tasks. Zephr Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Bill Chen, Founder Co-founder of Zephr. Born in Shanghai, I grew up mainly in Vancouver, Canada. I earned my bachelor's in CS from Columbia. Before Zephr, I worked at Retool (W17) as a Deployed Engineer. ### William Hu, Founder Building Zephr, an AI customer success manager. Previously @ Retool, Robinhood. ### Company Launches [### Zephr - Your AI customer success manager]( Hello YC! We are William and Bill, the team behind [Zephr]( TL;DR ----- * Zephr is an **AI customer success manager** that lives in Slack. Zephr automates manual tasks like writing emails, creating Linear tickets, and updating Salesforce records. * **Want a global view?** Zephr comes with a customer health dashboard and automatically identifies expansion opportunities and churn. * **Have a large team?** Motivate them with our CSM leaderboard. The Problem ----------- Customer success teams drive post-sale engagement, ensuring customer satisfaction, retention, and growth. Unlike customer support, customer success is proactive and contributes directly to revenue goals. However, the jobs of CSMs are operationally heavy, with most of their time spent on repetitive tasks rather than strategic initiatives. This is where Zephr comes in! Introducing Zephr! ------------------ Zephr is the first AI customer success manager. Zephr lives in Slack and automates routine tasks, such as writing emails, escalating tickets, and updating Salesforce records. With Zephr, you can manage 100 customers with ease. In addition to the Slack bot, Zephr comes with a lightweight customizable customer health dashboard. Zephr monitors customer health and automatically identifies expansion opportunities and risky accounts. Have a large team? Zephr provides a leaderboard to keep your team engaged. Our AI automates all the boring stuff, so CSMs can focus on what matters: being the face of the company and building relationships that drive measurable value. Zephr is not just making it easier for customer success teams—we're redefining what customer success can be. The Team -------- We met at a high school research camp a decade ago, became coworkers at Retool, and are now co-founders! Bill was a member of the customer success team at Retool, so he has experienced these challenges first-hand. William, on the other hand, was an engineer at Retool and enjoys solving these problems with software. The Ask 🙏 ---------- If you're setting up a B2B success motion or know someone who is, [come talk to us]( We'd love to see how our tool could fit into your workflow.

Poka Labs

poka-labs

https://www.pokalabs.com/

San Francisco, CA, USA

Poka Labs is developing an AI platform to automate operations tasks in chemical manufacturing, beginning with production scheduling. Traditionally, these tasks are performed manually using spreadsheets. Our software seamlessly integrates with existing data sources such as data historians, emails, and PDFs to automate analytics, scheduling, changes, and communication within a single platform. Malay and Andrew met while pursuing their MBAs at Harvard Business School. Malay previously worked as an engineer in the specialty chemical industry, saving his employers over $100 million across US, China, and German plants. Andrew worked as a software engineer on data infrastructure at Meta and optimization problems at a seed-stage startup. He also completed a Masters degree at Harvard specializing in Machine Learning.

The modern operating system for chemical manufacturing.

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/poka-labs

https://yc-oss.github.io/api/batches/s24/poka-labs.json

Title: Poka Labs: The modern operating system for chemical manufacturing. | Y Combinator URL Source: Markdown Content: ### The modern operating system for chemical manufacturing. Poka Labs is developing an AI platform to automate operations tasks in chemical manufacturing, beginning with production scheduling. Traditionally, these tasks are performed manually using spreadsheets. Our software seamlessly integrates with existing data sources such as data historians, emails, and PDFs to automate analytics, scheduling, changes, and communication within a single platform. Malay and Andrew met while pursuing their MBAs at Harvard Business School. Malay previously worked as an engineer in the specialty chemical industry, saving his employers over $100 million across US, China, and German plants. Andrew worked as a software engineer on data infrastructure at Meta and optimization problems at a seed-stage startup. He also completed a Masters degree at Harvard specializing in Machine Learning. ### Jobs at Poka Labs San Francisco, CA, US / New York, NY, US $140K - $160K 1.00% - 2.00% 3+ years Poka Labs Founded:2023 Team Size:2 Location:San Francisco ### Active Founders ### Andrew Bass, Founder Andrew is the Co-Founder and CTO of Poka Labs. Before Poka, he worked on critical data infrastructure systems at Meta, as well as optimization software at a Series A startup (acquired) that helped mobility companies boost their revenue 40%. He has both an MBA and MS degree at Harvard where he focused on ML systems. ### Malay Shah, Founder Malay is the Co-Founder and CEO of Poka Labs. He earned degrees in Chemical Engineering and Economics from NC State and an MBA from Harvard Business School. With a background in the specialty chemicals industry, Malay has held various positions in operations and engineering working across plants in US, Germany, and China. His work in deploying chemical expertise with advanced analytics has saved his previous employers tens of millions of dollars. ### Company Launches [### Poka Labs: The Modern Operating System for Chemical Manufacturing]( ### Summary: * We help chemical manufacturers automate their production planning and scheduling without the need for spreadsheets or complex implementations * Schedule a demo - [ Hi everyone! We’re Andrew and Malay — the founders of Poka Labs. ### ❌ Problem Production planning is the “brains” of chemical operations. Yet even today, the $5.6 trillion chemical industry relies on spreadsheets and humans to manage the process. This leaves plants to constantly fight fires in order to meet customer orders on time and maximize manufacturing margins. Any change, such as inventory delays, labor shortages, or unexpected maintenance, further stresses operations. Existing systems, such as ERPs, are too rigid in their data inputs to adapt to these types of changes. ### ✅ Our Solution: Adaptive Production Planning with AI Poka Labs integrates with existing information (data historians, emails, PDFs, ERPs, etc) to put chemical production scheduling on **autopilot**. Our software automates analytics, scheduling, changes, and communication within one platform that anyone can use. With Poka Labs, chemical plants can: * Maximize revenue * Minimize transition costs * Eliminate messy spreadsheets * Increase visibility ### 🎉 Our Story [Malay]( is a chemical engineer who has worked in chemical plants across 3 continents. He experienced these issues firsthand and saved his prior employers over $100+ million on process improvement and production planning projects. He met [Andrew]( while they were pursuing their MBA at Harvard Business School. Andrew worked as a software engineer on data infrastructure systems at Meta before working on optimization problems at a seed-stage startup. He also completed a Masters degree at Harvard specializing in ML. Together, we’re passionate about deploying modern software in the chemical industry. ### ❓ The Ask * If you work in chemical manufacturing, we’d love to chat! Contact [[email protected]](mailto:[email protected]) * Share this post! Is there anyone in your network that deals with planning for chemicals? We’d love to chat with them * _Copy & paste blurb_: A team of Harvard grads created an AI-powered production planning platform designed for the chemical industry. It automates scheduling, changes, and analytics in one platform. Put your plant on autopilot with Poka Labs. Contact [[email protected]](mailto:[email protected]) to see a demo from the founders.

Lilac Labs

lilac-labs

https://www.drive-thru.ai/

San Francisco, CA, USA

At Lilac, we automate the person taking order at the drive thru with a voice AI. We're building this for Quick Service Restaurants (QSRs) dealing with an historical labor shortage and rising wages. Previous attempts at drive-thru voice ordering are costly to implement and have failed to deliver the accuracy and latency needed. It's now possible to build a voice interface that passes the threshold for a great customer experience. In the United States, there are 200,000 Drive-Thrus handling 6 Billion visits a year. At 3 minutes per order, that's 34,000 human years spent on taking orders annually. Per location, on average we can deliver around $100,000 of value in terms of labor savings, upsell revenue lift, and training costs.

We automate the person taking orders at drive-thrus with a voice AI

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/lilac-labs

https://yc-oss.github.io/api/batches/s24/lilac-labs.json

Title: Lilac Labs: We automate the person taking orders at drive-thrus with a voice AI | Y Combinator URL Source: Markdown Content: ### We automate the person taking orders at drive-thrus with a voice AI At Lilac, we automate the person taking order at the drive thru with a voice AI. We're building this for Quick Service Restaurants (QSRs) dealing with an historical labor shortage and rising wages. Previous attempts at drive-thru voice ordering are costly to implement and have failed to deliver the accuracy and latency needed. It's now possible to build a voice interface that passes the threshold for a great customer experience. In the United States, there are 200,000 Drive-Thrus handling 6 Billion visits a year. At 3 minutes per order, that's 34,000 human years spent on taking orders annually. Per location, on average we can deliver around $100,000 of value in terms of labor savings, upsell revenue lift, and training costs. Lilac Labs Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Tony Kam, Co-Founder / CEO Tony is the co-founder and CEO of Lilac Labs. Tony studied EECS at UC Berkeley where focus on operating systems. Before Lilac, Tony worked at Tesla as an engineer, where he owned the display and touch component of the infotainment system. Tony also worked on the performance and reliability of driving visualization. Prior to that, Tony worked on manufacturing automation at Intel and robotics simulation at CloudMinds. ### Shelden Shi, Co-Founder / CTO Shelden is the Co-Founder and CTO of Lilac Labs. Shelden studied CS and CogSci at UC Berkeley, where he conducted research training ML models to predict emotions via audio data. Before Lilac, Shelden was the founding engineer at Symbolic, leading the development of a Fintech lending platform with ~$1B in AUM and 30k monthly active users. Prior to that, Shelden worked at Flatiron Health, an oncology data unicorn, where he trained and built the evaluation pipeline for cancer diagnostic models. ### Company Launches [### Lilac Labs: Drive-thru Order Taking with Voice AI]( **TLDR: Lilac Voice is an AI team member that takes orders at your drive-thru.** -------------------------------------------------------------------------------- ❌ **Problem:** **The quick service restaurant (QSR) industry is desperate for solutions to reduce labor costs.** ---------------------------------------------------------------------------------------------------------------- Grappling with rising wages, high staff turnover, and an unprecedented labor shortage, quick-service restaurants (QSRs) have increased prices to maintain profitability, but have ran into the limit of what consumers are willing to accept. Drive-thru’s makes up 70% of QSR revenue as a channel and is a significant opportunity for automation. In the U.S., there are 200,000 Drive-Thrus that handle 6 billion visits a year. ✅ **Solution: We automate the person taking orders at the drive thru with a voice AI.** --------------------------------------------------------------------------------------- Lilac Voice is a multi-lingual AI agent that takes orders from customer through the drive-thru speaker post and sends it directly into the kitchen. It’s perfectly trained, understands the menu, and can answer questions about ingredients and allergies. With Lilac Voice, QSR’s with drive-thrus can: 1) save labor cost of up to 1 full-time employee, 2) boost revenue through consistent and relevant upsell, 3) improve the customer experience with faster speed of service and higher order accuracy, and 4) improve staff retention! As an example: In CA, as of April 1st, the fast food minimum wage grew from $16 to $20. For a drive-thru operating 16 hours a day, 365 days a year, that’s $100,000+ in labor savings alone. **🙋Our Ask: How you can help.** -------------------------------- **Share this post!** We’re looking for more customers to onboard for this Summer! We’d love intros to franchisee operators, restaurant groups, or leadership at corporate chain. **Quick blurb to copy & paste:** Lilac Labs is offering free pilots for their voice AI drive-thru solution. The team consists of ex-Tesla engineers and Berkeley researchers. Learn more at [ For a demo and free pilot, contact [[email protected]](mailto:[email protected]). \- [Tony]( and [Shelden](

Remade

remade

https://www.remade.ai/

San Francisco, CA, USA

Remade uses AI to create studio-quality videos. For example, ScentBird, a fragrance subscription service, uses Remade to create TikTok ad hooks and increase conversion. Traditional videography is labour-intensive, slow, and expensive. Enterprises spend thousands on it, while smaller businesses often can't afford it. Remade reduces the cost of professional videos by 100x and cuts delivery time from weeks to 15 minutes. Using AI and social media data, we personalize content creation and optimize product ads. Our mission is to streamline visual content generation across multiple industries, making it faster and more accessible for businesses.

Automating Video Ad Workflows using AI

B2B

B2B -> Marketing

S24

Active

Early

https://www.ycombinator.com/companies/remade

https://yc-oss.github.io/api/batches/s24/remade.json

Title: Remade: Automating Video Ad Workflows using AI | Y Combinator URL Source: Markdown Content: ### Automating Video Ad Workflows using AI Remade uses AI to create studio-quality videos. For example, ScentBird, a fragrance subscription service, uses Remade to create engaging TikTok ad hooks and increase conversion. Traditional videography is labour-intensive, slow, and expensive. Enterprises spend thousands on it, while smaller businesses often can't afford it. Remade reduces the cost of professional videos by 100x and cuts delivery time from weeks to 15 minutes. Using AI and social media data, we personalize content creation and optimize product ads. Our mission is to streamline visual content generation across multiple industries, making it faster and more accessible for businesses. Remade Founded:2024 Team Size:4 Location:San Francisco ### Active Founders ### Alex Matthews, Founder Alex is the CEO of Remade, a company enabling businesses to automate their entire visual content marketing using AI. Remade's web application transforms basic, low quality photos of any product into professional hooks. Alex holds a BA in Computer Science from the University of Cambridge. He has contributed to earning £700k+ in revenue/ funding in past ventures. ### Christos Antonopoulos, Founder Christos Antonopoulos is the co-founder of Remade. Remade transforms basic, low quality photos of any product into professional advertisement videos . He holds a BA and MEng in Information & Computer Engineering from the University of Cambridge and has research contributions in Machine Learning for Medical Diagnosis and Natural Language Interfaces. ### Blendi Bylygbashi, Founder Blendi is the co-founder of Remade, a company focused on professional photography using AI. He holds a BA and an MEng in Information and Computer Engineering from the University of Cambridge. At Transport for London, he developed deep learning models for CCTV event detection. His research on intra-cranial hypertension was published in the Lancet Neurology Journal. He also managed the BlenDigi YouTube channel, growing it to almost 1 million subscribers and generating $200,000 in ad revenue. ### Rehan Sheikh, Founder Rehan is the CTO and co-founder of Remade, a company focused on professional photography using diffusion models. He holds a BA and MEng in Information & Computer Engineering from the University of Cambridge! ### Company Launches [### Remade: Automating video ad workflows using AI]( [Launch Video]( **Tl;DR: Remade automates visual marketing workflows. Lifestyle product brands generate TikTok hooks that land 130% higher clickthrough rate.** **Try it now:** [** Hi everyone! We're Alex, Rehan, Chris, and Blendi — the founders of [Remade]( We are excited to announce our product image-to-video workflow, enabling the transformation of low-quality product images into video hooks that sell. How it works: ------------- 1. **Upload an image of your product** 2. **Use “AI Backgrounds” to create a scene for your product video.** 3. **Generate your TikTok hook** [TikTok Hook]( Use Cases --------- **Marketing Teams** Marketing teams use Remade to generate short-form product reels in minutes. We replace a process that corporations outsource to marketing agencies for thousands of dollars in 3-clicks. **Marketplaces** Marketplaces use Remade’s Enterprise API to quickly generate product videos for their catalogs. These videos help sellers showcase products with dynamic, attention-grabbing content, which is proven to boost conversion, reduce bounce rate, and enhance the shopping experience. **Delivery Platforms** Delivery platforms use Remade to generate videos for their food listings, enabling a TikTok-style User Experience. **👋**Asks: How you can help ---------------------------- Warm intros to Heads of Image Catalogue at Delivery platforms such as Zomato, Uber Eats, Deliveroo, Glovo and Grubhub. Warm Intros to CMO’s / Heads of Content of lifestyle product brands. Email us at [[email protected]](mailto:[email protected]) for 24/7 support. **Try it here:** [** ### Other Company Launches ### Remade: GenAI product photography for businesses Generate studio-quality product photoshoots from your smartphone. [Read Launch ›]( #### YC Sign Photo

XTraffic

xtraffic

https://XTraffic.com/

Dallas, TX, USA

XTraffic is at the intersection of advancements in sensor technology and affordability. Where it would previously cost a city up to $100,000 per intersection to create intelligent traffic lights, we can accomplish the same for pennies on the dollar. This means entire cities can affordably upgrade their infrastructure - and some already are. We are live in multiple Texas cities, with both ongoing and successful pilots, corridors of multiple intersections, and soon to be entire cities. Our customers enjoy less traffic, better safety, and data on which to build their future.

Technology for cities to automate and manage their traffic lights.

Government

Government

S24

Active

Early

https://www.ycombinator.com/companies/xtraffic

https://yc-oss.github.io/api/batches/s24/xtraffic.json

Title: XTraffic: Technology for cities to automate and manage their traffic lights. | Y Combinator URL Source: Markdown Content: ### Technology for cities to automate and manage their traffic lights. XTraffic is at the intersection of advancements in sensor technology and affordability. Where it would previously cost a city up to $100,000 per intersection to create intelligent traffic lights, we can accomplish the same for pennies on the dollar. This means entire cities can affordably upgrade their infrastructure - and some already are. We are live in multiple Texas cities, with both ongoing and successful pilots, corridors of multiple intersections, and soon to be entire cities. Our customers enjoy less traffic, better safety, and data on which to build their future. XTraffic Founded:2022 Team Size:4 Location:Dallas, TX ### Active Founders ### Luke Adams, Founder Aerospace Engineering + Computer Science. 3 years @ SpaceX wearing many hats on the Starship program. Now, working to bring traffic and city infrastructure into the modern age @ XTraffic! ### Everett Ivy, Founder Hi! I have studied computer science, worked on control systems at Amazon, and been a professional gamer. Now with my friends and cofounders at XTraffic I am making traffic lights smarter. ### Brian Payne, Founder Background in control systems and embedded hardware at Amazon. Love all things robotics, now building hardware and software systems to solve traffic! ### Company Launches [### XTraffic - Making traffic lights smarter and travel times faster]( **Tl;dr:** [XTraffic]( upgrades traffic lights to work together as an intelligent system, communicating and coordinating to reduce traffic and improve safety. The autonomous control of traffic lights gives back to both city officials and their citizens their most valuable resource - time. — Hi everyone, we’re [Everett]( [Luke]( and [Brian]( the founders of XTraffic. **Why Now?** The combination of rapidly advancing technology and smart-city initiatives has created a unique intersection in time where intelligent traffic control is not just possible but affordable. As smart city initiatives are being adopted worldwide, municipalities of all sizes are looking for ways to affordably transition to intelligent infrastructure. XTraffic leverages this trend by providing cities with a scalable, intelligent system that upgrades existing traffic lights into a self-optimizing network. **Market Opportunity** The global smart city market is projected to reach [$3.84 trillion by 2029]( with traffic management as a requirement to aid in driving this growth. Urban areas are expanding, and traffic congestion is a major pain point that costs the U.S. [more than $70.4 billion in 2023]( a 15% increase from 2022. XTraffic is tapping into this immense market opportunity by offering a solution that not only reduces traffic congestion but also enhances safety, reduces emissions, and increases overall quality of life. **Why XTraffic?** Our system isn’t just about less traffic; it’s about smarter use of our infrastructure. XTraffic's approach allows traffic lights to 'talk' to each other, making real-time decisions based on current conditions across an entire network. This holistic view enables our system to adapt dynamically to changing traffic patterns, emergencies, and other factors - this significantly reduces delays and optimizes traffic flow across entire cities. With our solution, cities can avoid the massive costs of overhauling infrastructure. Instead, they can implement an affordable, flexible system that integrates seamlessly with existing technology. This reduces the barrier to entry and allows for rapid deployment, scaling from single intersections to entire metropolitan areas. **Call to Action** We’re excited to be part of Y Combinator, and we’re looking forward to connecting with cities, investors, and partners who share our vision for smarter, safer, and more efficient urban environments. If you’re interested in learning more or discussing opportunities, check out [ or reach out directly to our team. Let’s work together to give people back their most valuable resource: time.

camfer

camfer

https://www.camfer.dev/

San Francisco, CA, USA

We’re building the first AI mechanical engineer that collaborates with humans to do design tasks end-to-end. Human engineers can talk to Camfer to build, test, and iterate 3D designs natively on CAD platforms.

Building the world’s first Al mechanical engineer.

B2B

B2B -> Engineering, Product and Design

S24

Active

Early

https://www.ycombinator.com/companies/camfer

https://yc-oss.github.io/api/batches/s24/camfer.json

Title: camfer: Building the world’s first Al mechanical engineer. | Y Combinator URL Source: Markdown Content: ### Building the world’s first Al mechanical engineer. We’re building the first AI mechanical engineer that collaborates with humans to do design tasks end-to-end. Human engineers can talk to Camfer to build, test, and iterate 3D designs natively on CAD platforms. camfer Founded:2024 Team Size:3 Location:San Francisco ### Active Founders ### Arya Bastani, Founder/CEO Before working at AWS and graduating from UC Berkeley in 2023, I was president of my high school FRC robotics team - ranked 4th in the world. This is where I met Roth Vann. While studying CS @ Berkeley: I was Chief EECS Engineer on the Formula Electric Racing team, I conducted research in Professor Anant Sahai’s lab doing brainwave (EEG) classification and generation using transformers and stable diffusion, and was President of the Iranian Students Cultural Org. More @ aryabastani.com ### Keaton Elvins, Founder EECS honors grad @ UC Berkeley, helped launch Amazon Q to millions of users ### Roth Vann, Founder Dropped out of UCR to work on ads at Meta. Dropped out of Meta to build Camfer.

Henry

henry-2

http://www.henry.ai

New York, NY, USA

Henry is an AI copilot for commercial real estate (CRE) brokers that seamlessly integrates a brokerage’s internal data set with external sources to generate custom presentations and financial modeling for deals. Our mission is to help CRE brokers close more deals faster, earning more while doing less repetitive work. We’re initially focusing on enabling brokers to generate deal decks in seconds—a task that typically consumes 20+ hours a week across multiple departments within a brokerage.

Generative AI for Commercial Real Estate Professionals

Real Estate and Construction

Real Estate and Construction

S24

Active

Early

https://www.ycombinator.com/companies/henry-2

https://yc-oss.github.io/api/batches/s24/henry-2.json

Title: Henry: Automating Deal Decks for Commercial Real Estate Brokers | Y Combinator URL Source: Markdown Content: ### Automating Deal Decks for Commercial Real Estate Brokers Henry is an AI copilot for commercial real estate (CRE) brokers that seamlessly integrates a brokerage’s internal data set with external sources to generate custom presentations and financial modeling for deals. Our mission is to help CRE brokers close more deals faster, earning more while doing less repetitive work. We’re initially focusing on enabling brokers to generate deal decks in seconds—a task that typically consumes 20+ hours a week across multiple departments within a brokerage. Henry Founded:2024 Team Size:3 Location:New York ### Active Founders ### Sammy Greenwall, Founder, CEO Sammy is the Co-Founder and CEO of Henry, Co-Founder and former CRO of Lev (which he scaled 0 to Series B / $10M+ run rate), and a retired real estate finance professional. Sammy was born and raised in the Bay Area and played college basketball at Swarthmore College. Sammy is passionate about building great products in large and stubborn markets. ### Adam Pratt, Founder 2x Founder, Currently CTO and Co-Founder @ Henry. Previously, Firefighter (Hamilton, NY - Station 19) and Co-Founder @ Halligan (Fire Department Saas Platform). Aquired by Vector Solutions in 2019 ### Company Launches [### Henry - AI copilot for commercial real estate brokers]( Hi everyone! We are Sammy and Adam, and we’re creating an AI copilot that automates the deal deck creation and financial analysis process for CRE (commercial real estate) brokers. For example, [Henry]( can create a full deal package for a broker selling an office building in Dogpatch in minutes instead of weeks. This will ensure that brokers maximize their time in the most value-added part of their business: relationships and sales. Traditional brokers often spend as much as 50% of their time on tasks that do not contribute to their bottom line. From basic financial analysis to marketing presentations to win future business to basic operational support, brokers often struggle with the manual and time-consuming realities of their business. These tasks are an incredibly painful bottleneck for brokers, which interferes with their primary goal of _building relationships to buy and sell property_. We are creating [Henry]( to solve this problem. Henry is an AI copilot that takes all of the messy, unstructured data that defines CRE transactions and consolidates it into clean marketing deliverables & financial analyses in minutes rather than weeks. Long term, we plan on automating the entire deal flow process for brokers to make their cycle time for a deal an order of magnitude faster. You can find an example of what we do below: [ I ([Sammy]( led Lev for the last five years, a 200+ person series B real estate financing marketplace that I co-founded. During this time, I realized that brokers were often bottlenecked by manual processes & antiquated technology, a problem optimally solved with advances in AI-driven technology. This led me to my partnership with [Adam]( who has deep experience in creating AI-enabled solutions for thorny problems that plague traditional businesses. **How you can help:** --------------------- * Know any CRE brokers? We’d love to talk to them about how we can make their lives easier & their paychecks larger. Contact us at [[email protected]](mailto:[email protected])

ACX

acx

https://www.acxtherapeutics.com/

San Francisco, CA, USA

ACX is the only company that has successfully reproduced the compounds that bacteria use to kill microbes for therapeutic development. Our patentable technology, which comprises synthetic compounds that mimic natural substances, has applications in humans, animals, and crops. We are beginning with the elimination of crop pathogens—harmful to pests but beneficial for humans!

Therapeutics inspired by bacteria's natural killing abilities

Healthcare

Healthcare -> Industrial Bio

S24

Active

Early

https://www.ycombinator.com/companies/acx

https://yc-oss.github.io/api/batches/s24/acx.json

Title: ACX: Therapeutics inspired by bacteria's natural killing abilities | Y Combinator URL Source: Markdown Content: ### Therapeutics inspired by bacteria's natural killing abilities ACX is the only company that has successfully recreated synthetically the compounds that bacteria use to kill microbes for therapeutic development. Our patentable technology has applications in humans, animals, and crops. We are beginning with the elimination of crop pathogens—harmful to pests but beneficial for humans! ACX Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Emmanouela Petsolari, Founder Emmanouela is the Co-Founder & CEO of ACX and PhD in Biochemistry and Biophysics from the University of Cambridge. Previously, she held several research positions working in immunology, cancer, malaria vaccine development and viral drug discovery (in institutes like Barts Cancer Institute CRUK, Cambridge University, King’s College London, Institute Curie France and Imperial College London). ### Melina Petsolari, Founder Melina is the Co-founder & CTO of ACX and PhD in Computer Science from King's College London. Previously, she held several research positions spanning digital health, robotics, technology-enable interventions, HCAI and NLP (in institutes like Cambridge University, King’s College London, National Health Service (NHS) in England and Francis Crick Institute). ### Company Launches [### 🧬🦠 ACX – Therapeutics inspired by bacteria's natural killing abilities]( Hi everyone, We are [**ACX**]( We’re two PhD’s in Biochemistry/Biophysics and Computer Science from the University of Cambridge and King’s College London developing a new type of therapy with applications in everything from cancer to agriculture. We have discovered a new molecular weapon that naturally occurs in bacteria to eliminate bad microbes with high specificity. We are harnessing this natural killing mechanism to first provide more efficacious pest control measures. ### **Team:** 👩🏼‍🔬[**Emmanouela Petsolari**]( – Molecular Geneticist and Biochemist with 8 years of research experience in drug discovery and development (from leukemia, ovarian, and breast cancer as well as viral and bacterial infections). 👩🏼‍🔬 [**Melina Petsolari**](  – Computer Scientist and Designer with extensive experience in HCI, AI, and healthcare developing hardware and software for applications in digital health, biotech, and privacy. **❌ Problem:** -------------- ### **_In humans_** Despite billions of dollars being spent each year on cancer therapeutics and cancer research, more than 10 million patients still die every year from the disease. 1. ~90% of cancer-related deaths in patients receiving chemotherapy are caused due to ineffective treatments that have off-target effects ●  Cancer relapse and metastasis formation further lead to less than five-year survival rates ●  Current treatments cause high toxicity and adverse side effects due to a lack of on-target specificity, which further causes disruption of healthy cells ### **_In animal farming and agriculture_** There is more disease impacting our livestock and crops than ever before - and this comes at a time when the human population is also exploding. 1. Manure from livestock and extensive use of limited therapeutics for growth promotion and disease prevention in animal breeding as well as overuse of pesticides in crops ●  One billion kilos of chemicals are used annually to combat crop pests 2. Spread of pathogenic agents that have adapted in previously used medications 3. Failure of commonly used pesticides and insecticides has been documented in nearly 1000 distinct pest species worldwide ●  This leads each year to the world’s food supply facing an up to 30% loss due to insect pests alone, and this number is continuously increasing ✅ **Our Solution:** ------------------- What do these three areas have in common? Existing therapeutics are either no longer effective or have never worked to the extent necessary for achieving positive outcomes. We have discovered a common biological pathway that can be targeted from the human level down to the animal level and even to the crop level to develop robust therapeutics capable of overcoming pathogen and cancer evolution. ### ✅ ✅ **Specifics:** ACX is the only company that has been able to reproduce the mechanism that bacteria use to kill microbes for therapeutics, and we are starting by eliminating crop pathogens: 1. We have _in vitro_ validation data of our therapeutic showing high inhibition and specificity in the cell line 2. We are initially focusing on animal health and crop production due to the experimental validation of our therapeutic's effectiveness on a shared pathogen 3. We are developing drug delivery systems that minimize toxicity and immunogenicity while providing precise spatiotemporal control for targeted drug release 4. Considering the human case, our therapeutic has the potential to treat various cancer types and become a standard-of-care treatment with blockbuster potential Illustration of our solution. Nanoparticles carrying the killer compound (orange) infiltrate the target cell like a Trojan horse and activate drug release upon targeting our newly discovered pathway. ### 💡**Key facts about our solution:** ➔   High killing efficacy ➔   Designed to specifically target malignant cells while preserving the surrounding healthy cells and ensuring no toxicity or adverse side effects ➔   Highly suitable to become a platform solution (i.e., therapeutic development pipeline expanded to multiple disease areas and organisms) ➔   Scalable, stable, and easy to manufacture 👋**Our Asks:** --------------- * Share this post with your network and help spread the word! * If you’re as excited as we are about this novel approach, email us to learn more about it at [[email protected]](mailto:[email protected])

ClaimSorted

claimsorted

https://claimsorted.com

ClaimSorted helps insurance companies remove the hassle of managing claims by enabling them to outsource their claim operations to us. This service is known as a Third Party Administrator (TPA). Unlike traditional TPAs, we blend AI and best-in-class experts to deliver a 5-star customer experience, minimise mistakes and speed up claim assessment.

Making it easy for insurance companies to process claims

Fintech

Fintech -> Insurance

S24

Active

Early

https://www.ycombinator.com/companies/claimsorted

https://yc-oss.github.io/api/batches/s24/claimsorted.json

Title: ClaimSorted: Making it easy for insurance companies to process claims | Y Combinator URL Source: Markdown Content: ### Making it easy for insurance companies to process claims ClaimSorted helps insurance companies remove the hassle of managing claims by enabling them to outsource their claim operations to us. This service is known as a Third Party Administrator (TPA). Unlike traditional TPAs, we blend AI and best-in-class experts to deliver a 5-star customer experience, minimise mistakes and speed up claim assessment. ### Jobs at ClaimSorted London, England, GB £86K - £132K GBP 0.50% - 1.00% 3+ years ClaimSorted Founded:2024 Team Size:2 Location: ### Active Founders ### Pavel Gertsberg, Founder Before starting ClaimSorted, Pavel built a pet insurance company, Fluffy, which he scaled to over 20,000 customers working with some of the largest insurers. Before starting his insurance company, Pavel worked as a Head of Growth building marketing, sales and product functions in SaaS and InsurTech companies. ### German Mikulski, Founder Co-Founder and CTO at ClaimSorted. Previously founded Fluffy, a pet insurance company, automating ~70% of back-office processes with AI. Formerly at Deutsche Bank, where he developed Big Data systems processing over $100 million in transactions ### Company Launches [### ClaimSorted: Hassle-free insurance claims with AI]( **TL:DR: We help insurers manage claims efficiently by outsourcing claim operations to us. We then use AI to make the process faster and cheaper.** **🙋 Ask: looking for intros to insurance companies.** ### **Problem** If you've ever submitted an insurance claim and it took ages to get your money back, here's why: **Insurers outsource claims** to claim outsourcing agencies, known as **Third Party Administrators (TPAs)**. Imagine a huge warehouse with **thousands of people using pen and paper** to process millions of claims. That’s why it's so slow. ### **Solution** At [ClaimSorted]( we are an **AI-first TPA for insurance claims**. We automate fraud checks, compliance, claim decision-making, and **deliver payouts in minutes**. Our AI is paired with a team of experts to ensure accuracy. ### **Team** We **previously built an insurance company** and know the struggles of working with TPAs. We aim to make every claim a positive experience for customers and drive cost efficiencies for insurers. _Fun fact, _[_German_]( was the best man at _[_Pavel_]( wedding_ ### **Ask:** **If you know any insurance companies in the US, UK, or EU, please email us at [[email protected]](mailto:[email protected])**

RowBoat Labs

rowboat-labs

https://www.rowboatlabs.com

San Francisco, CA, USA

RowBoat Labs offers pre-trained LLM agents for customer support, which continuously learn from usage. Our LLMs are safe and brand-aligned. RowBoat agents seamlessly integrate into your systems and take actions where necessary.

LLM Agents for Customer Support

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/rowboat-labs

https://yc-oss.github.io/api/batches/s24/rowboat-labs.json

Title: RowBoat Labs: LLM agents for customer support in fintech | Y Combinator URL Source: Markdown Content: ### LLM agents for customer support in fintech RowBoat Labs offers pre-trained LLM agents for customer support, which continuously learn from usage. Our LLMs are safe and brand-aligned. RowBoat agents seamlessly integrate into your systems and take actions where necessary. RowBoat Labs Founded:2024 Team Size:3 Location:San Francisco ### Active Founders ### Arjun Maheswaran, Founder I previously co-founded and served as CTO of Agara, a customer support AI startup that was acquired by Coinbase in 2021. Over the past decade, I've focused primarily on Deep Learning for Natural Language Processing (NLP), at Twitter Cortex, Agara, and Coinbase. ### Ramnique Singh, Founder Co-founder at RowBoat Labs. Helping companies supercharge CX using LLMs. ### Akhilesh Sudhakar, Founder Building LLM agents for customer support at RowBoat Labs. Most recently, led generative AI product for CX at Coinbase. Previously, ML scientist at Agara AI (autonomous customer support). ### Company Launches [### RowBoat Labs - LLM agents for human-like customer support]( **Tl;dr:** [RowBoat]( is a pre-trained customer support LLM agent that seamlessly plugs into your systems, handles customer conversations, and performs tasks just like an expert human agent would. **Problem** ----------- Customer support is one of those rare opportunities for a brand to connect directly with its users. The best brands know this and invest heavily in coaching their human agents to deliver a great support experience. But when it comes to automated support, most solutions rely on LLMs that aren’t specially designed for customer support and can’t learn from experience. This often leaves users stuck with impersonal and long FAQ-style responses in small chat windows, often missing the help they actually need. Internal teams might see some early wins with a public LLM + RAG setup, but that progress usually hits a wall fast. Before long, engineers are spending more time debugging common LLM issues instead of focusing on what actually matters: delivering great customer support. As a result, customer satisfaction rates (measured by surveys) get nowhere close to that of human agents. **Solution** ------------ RowBoat is built and fine-tuned specifically for customer support * **Pre-Trained:** RowBoat is trained on a vast corpus of support conversations and further refined through self-play. It's like hiring a highly experienced customer support agent right from day one. * **Brand-Aligned:** RowBoat automatically indexes your public and internal knowledge, fine-tuning itself to align with your brand. * **Continuous Improvement:** RowBoat learns from every user interaction. When connected to your user metrics, such as resolution rates, RowBoat critiques its own performance to optimize for them continuously. … for internal engineering teams to build exceptional customer experience * **Plug & Play:** A drop-in replacement for GPT-4+ class LLMs, RowBoat offers immediate improvements in resolution accuracy. Our SmartRAG system readily integrates with your existing Elasticsearch or embedding-based retrieval systems for improved grounding. * **Tool Use & Personalization:** Equipped with a library of predefined functions, RowBoat can interact with and perform tasks on internal and external APIs. This makes RowBoat’s conversations highly contextualized to the user. * **Self-hostable:** RowBoat can be hosted inside your company’s cloud, especially for privacy-sensitive use cases. Seamless conversations… … with continuous coaching **Who we are** -------------- We're a team of three engineers who have worked together for the last 7 years. [Arjun]( co-founded Agara, a customer support AI startup where [Ramnique]( and [Akhilesh]( were part of the founding team. In 2021, Agara was acquired by Coinbase, where we built their automated customer support. We have published research and hold patents on LLMs, Reinforcement Learning, embeddings, and customer support. **Our Ask** ----------- If your company has a customer support function, we’d love to connect with your CX or engineering team. Please reach out to us at [[email protected]](mailto:[email protected]). We’ll bring an LLM agent trained for your brand to our call 😊

Olive Legal

olive-legal

https://olive.legal

San Francisco, CA, USA

Olive uses AI to summarize client medical records for personal injury lawyers. We're at $5k MRR after launching five weeks ago, and law firms choose us because we double paralegal efficiency. Sam's ex-girlfriend introduced him to Greg back at CMU in 2017, and while that relationship didn't last, their friendship has. After undergrad, Greg went to Harvard Law School, while Sam worked for three years at Jane Street, including a year in Hong Kong where he built out a satellite dev team for the algo options trading desk. Greg graduated, Sam quit, and we founded Olive with a big idea: use AI to make the law more accessible. We started with AI in corporate law given Greg's background, but quickly realized the space was crowded, and turned our attention to the under-competed $65B personal injury market. AI disruption makes a lot of sense in personal injury because the incentives are aligned—plaintiff lawyers are paid on contingency, and therefore love time-saving tools. We have competitors who have proven substantial demand for medical chronologies, but they operate at best on hybrid human/AI approaches with multi-day turnarounds. We think there's space for an AI solution that cuts humans out of the loop entirely with instant turnarounds, and we think we're the right team to do it. Up next, we will capture an increasing share of the value paralegals provide personal injury lawyers. We also see Olive expanding beyond just personal injury—interpreting unstructured medical data is critical for insurers to defend malpractice claims, class action firms to assemble plaintiffs, claimants to appeal insurance denials, providers to improve outcomes, the Social Security Administration to process disability applications speedily—we could go on. We're on a mission to remove barriers to information transfer, and therefore make justice more accessible.

AI medical summaries for personal injury lawyers.

B2B

B2B -> Legal

S24

Active

Early

https://www.ycombinator.com/companies/olive-legal

https://yc-oss.github.io/api/batches/s24/olive-legal.json

Title: Olive Legal: AI medical summaries for personal injury lawyers. | Y Combinator URL Source: Markdown Content: ### AI medical summaries for personal injury lawyers. Olive uses AI to summarize client medical records for personal injury lawyers. We're at $13k MRR after launching eight weeks ago, and law firms choose us because we double paralegal efficiency. Sam's ex-girlfriend introduced him to Greg back at CMU in 2017, and while that relationship didn't last, their friendship has. After undergrad, Greg went to Harvard Law School, while Sam worked for three years at Jane Street, including a year in Hong Kong where he built out a satellite dev team for the algo options trading desk. Greg graduated, Sam quit, and we founded Olive with a big idea: use AI to make the law more accessible. We started with AI in corporate law given Greg's background, but quickly realized the space was crowded, and turned our attention to the under-competed $65B personal injury market. AI disruption makes a lot of sense in personal injury because the incentives are aligned—plaintiff lawyers are paid on contingency, and therefore love time-saving tools. We have competitors who have proven substantial demand for medical chronologies, but they operate at best on hybrid human/AI approaches with multi-day turnarounds. We think there's space for an AI solution that cuts humans out of the loop entirely with instant turnarounds, and we think we're the right team to do it. We're focusing on the medical malpractice niche within personal injury, because there's a massive access to justice problem: medical malpractice lawyers decline most cases under $250k since they're too expensive to litigate. An estimated 80% of medical malpractice victims can't secure representation right now, and we're on a mission to change that by making the litigation process cheaper, at the same time unlocking a massive market of latent demand. Olive Legal Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Greg Volynsky, Founder Co-founder and CEO of Olive. Greg graduated Harvard Law School cum laude, where he focused on administrative law, comparative electoral systems, and avoiding contract law. Greg was a summer associate at Cravath. He bootstrapped his first company, which continues to operate, as a freshman at Carnegie Mellon. Greg was a fellow at BRI Excel Ventures. Greg is interested in 20th century political history, Soviet bard music & climbing. ### Sam Damashek, Founder Sam is the co-founder and CTO of Olive. He worked for three years at Jane Street on the options trading desk writing algorithmic strategies using applied ML, for two years in NYC and then one year in Hong Kong, where he built out a satellite dev team for the Asia options markets. He graduated with a BS in Computer Science from CMU, where for some reason he led a fledging constitutional law debate team, never to victory though frequently to Ohio. ### Company Launches [### Olive Legal: AI summaries of medical records for personal injury lawyers]( **tl;dr:** Olive uses AI to summarize client medical records for personal injury lawyers. We're at $5k MRR after launching five weeks ago, and PI firms choose us because we double the efficiency of their paralegals. Hi everyone! We’re [Sam]( and [Greg]( and we’re the team behind [Olive]( Sam's ex-girlfriend introduced him to Greg back at CMU in 2017, and while that relationship didn't last, our friendship has. After undergrad, Greg (left) went to Harvard Law School ⚖️, while Sam (right) worked for three years at Jane Street, including a year in Hong Kong, where he built out a satellite dev team for the algo options trading desk 📈. Greg graduated, Sam quit, and we founded Olive with a big idea: use AI to make the law more accessible. Given Greg's background, we started with AI in corporate law but quickly realized the space was crowded and turned our attention to the under-competed $65B personal injury market. Why personal injury law? ------------------------ The incentives are aligned—plaintiff lawyers are paid on contingency and, therefore, love time-saving tools. Paralegals are expensive, and LLMs are getting good enough that, with care, they can replace specific paralegal tasks. Why Olive? ---------- We have competitors who have proven substantial demand for medical chronologies, but they operate at best on hybrid human/AI approaches with multi-day turnarounds. We think there's space for an AI solution that cuts humans out of the loop entirely with instant turnarounds, and we think we're the right team to do it. Okay, but actually, why call it Olive 🫒? ----------------------------------------- I guess we liked the color scheme? What’s next? ------------ We will capture an increasing share of the value paralegals provide personal injury lawyers. We also see Olive expanding beyond just personal injury—interpreting unstructured medical data is critical for insurers to defend malpractice claims, class action firms to assemble plaintiffs, claimants to appeal insurance denials, providers to improve outcomes, the Social Security Administration to process disability applications speedily—we could go on. We're on a mission to remove barriers to information transfer, and therefore make justice more accessible. Our ask 🙏 ---------- If you know any personal injury lawyers, send them over to [our website]( or have them [book a demo directly](

Simplex

simplex

https://simplex.sh

San Francisco, CA, USA

Simplex creates on-demand vision datasets rendered from 3D scenes to train AI models. We can create data for any scenario, saving companies millions of hours they’d otherwise spend collecting and labeling real data.

Synthetic datasets for vision models

B2B

B2B -> Engineering, Product and Design

S24

Active

Early

https://www.ycombinator.com/companies/simplex

https://yc-oss.github.io/api/batches/s24/simplex.json

Title: Simplex: Synthetic datasets for vision models | Y Combinator URL Source: Markdown Content: ### Synthetic datasets for vision models Simplex creates on-demand vision datasets rendered from 3D scenes to train AI models. We can create data for any scenario, saving companies millions of hours they’d otherwise spend collecting and labeling real data. Simplex Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Shreya Karpoor, Founder Shreya is the co-founder and CEO of Simplex. She holds a BS and MEng in Electrical Engineering and Computer Science from MIT. She previously built software at Tesla and Viam and researched locomotion and dexterous robotic manipulation at MIT. ### Marco Nocito, Founder Marco is the co-founder and CTO of Simplex. He holds a BS and MEng in Computer Science from MIT. He previously built machine learning models to generate synthetic data at Waymo and built data infrastructure tooling at Viam and Bloomberg. ### Company Launches [### Simplex: on-demand photorealistic vision datasets]( **TL:DR;** Simplex creates photorealistic vision datasets rendered from 3D scenes for AI model training. Submit a request [on our website]( to receive high-quality data and labels. **Data request for the above sample:** _“Generate images and labels of a home kitchen with household objects on a center table. I need a variety of household objects in a variety of lighting conditions. Our desired labels are semantic segmentation and depth maps.”_ Hi everyone, we’re [**Shreya**]( and [**Marco**]( two MIT grads building Simplex. Collecting vision data for model training is time-consuming, costly, and often unsafe. Shreya spent over 200 hours physically operating a robot to collect image training data during her research at MIT. Marco worked on machine learning for synthetic data at Waymo to solve this exact problem. We realized data scarcity wasn’t just an issue in robotics – it affects any company training vision models. When fine-tuning foundation models or building a new dataset from scratch, teams must curate existing data or label and collect data themselves. We resolve the data scarcity problem by generating photorealistic ground truth labeled datasets for **any scenario**. We can generate **millions of varied images** **from 3D scenes** using our physics engine pipeline. Here’s how you’d use Simplex: 1. Fill out our data request form [here]( – it takes less than a minute. 2. Give us feedback on a few sample image/label pairs that we generate. Repeat if necessary. 3. Once you’re satisfied, download your complete dataset. We support semantic segmentation, captions, simulated LiDAR, depth maps, and bounding boxes. You can generate large volumes of randomized scenes or provide a CAD/phone scan model for more specific scenes. **Our Ask** ----------- * If you or someone you know needs vision data, fill out our 30-second data request [form.]( We’re taking a limited number of early customers. * If you have a more complicated request or would otherwise like to contact us, email [[email protected]]( **The Team** ------------ [**Shreya**]( Computer science (BS and MEng) at MIT, software engineer at Tesla and Viam. Built simulation pipelines for locomotion and dexterous manipulation research at MIT.  [**Marco**]( Computer science (BS and MEng) at MIT, software engineer at Waymo, Bloomberg, and Viam. Built machine learning models to generate synthetic data at Waymo. #### YC Sign Photo

expand.ai

expand-ai

https://expand.ai/

San Francisco, CA, USA

expand.ai instantly turns any website into a type-safe API you can rely on. You can either request data from any website instantly or let expand.ai build up datasets for you. We take care of the hard parts like dealing with bot protection, scaling browser infrastructure and making sure that we only extract verified, correct information.

Turn any website into an API.

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/expand-ai

https://yc-oss.github.io/api/batches/s24/expand-ai.json

Title: expand.ai: Turn any website into an API. | Y Combinator URL Source: Markdown Content: ### Turn any website into an API. expand.ai instantly turns any website into a type-safe API you can rely on. You can either request data from any website instantly or let expand.ai build up datasets for you. We take care of the hard parts like dealing with bot protection, scaling browser infrastructure and making sure that we only extract verified, correct information. expand.ai Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Tim Suchanek, Founder Having been the first engineer at Prisma and founder of Stellate, I'm passionate about databases, schemas and APIs. Now at expand.ai, we're turning the web into a type-safe API. #### YC Sign Photo ### Hear from the founders #### How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?) Tim has been working on developer tools for the last 8 years, being founding engineer at [Prisma]( and then founding [Stellate]( building AI apps, he found that oftentimes getting the right data was the blocker to get started with the project.That’s why he’s building [expand.ai]( - to enable anyone to use the web as a data source to power their AI apps.

Manaflow

manaflow

https://manaflow.ai

San Francisco, CA, USA

Manaflow is a simple way to automate repetitive office work in tables. With one click, you can execute millions of tasks involving data retrieval, gluing APIs, and taking actions side-by-side with your office work co-pilot.

Automate repetitive office work in tables with AI

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/manaflow

https://yc-oss.github.io/api/batches/s24/manaflow.json

Title: Manaflow: AI Workflow Builder for Businesses | Y Combinator URL Source: Markdown Content: Manaflow: AI Workflow Builder for Businesses | Y Combinator =============== []( "Y Combinator") [About]( [What Happens at YC?]( Interview Guide]( Blog]( [Companies]( [Startup Directory]( Companies]( Directory]( YC]( [Startup Jobs]( [All Jobs]( Engineering]( Operations]( Marketing]( Sales]( Internship Program 2024]( Job Guide]( Coaching]( Startup Jobs Blog]( [Find a Co-Founder]( [Library]( [SAFE]( [Resources]( [Startup School]( Investors]( News]( Open main menu Apply for **W2025** batch.[Apply]( "Apply for W2025 batch.") [Home]( Manaflow ======== AI Workflow Builder for Businesses [S24]( Active [developer-tools]( Francisco]( * * * [Company]( [Jobs](  [ * * * ### AI Workflow Builder for Businesses Manaflow automates repetitive internal workflows involving data, APIs, and actions. Manaflow provides the infrastructure for businesses to build useful AI agents that can schedule recurring tasks, loop in human approval, and call custom tools. Instead of operations teams manually using internal tools, Manaflow agents listen to English instructions and operate tools in the background, interfacing directly with sources of truth, applications, and databases. Manaflow Founded:2024 Team Size:3 Location:San Francisco Group Partner:[Dalton Caldwell](  []( "LinkedIn profile") []( "Twitter account") ### Active Founders ### Austin Wang, Founder Co-founder & CEO of Manaflow / previously Google, NASA JPL, & Chess.com / Yale Physics / featured on Business Insider & The Economist Austin Wang [Manaflow](  []( "Twitter account") []( "LinkedIn profile") ### Lawrence Chen, Founder Co-founder of Manaflow / prev. Minion AI / Berkeley '24 Lawrence Chen [Manaflow](  []( "Twitter account") []( "LinkedIn profile") ### Wesley Tjangnaka, Founder Co-founder of Manaflow / Stanford CS Wesley Tjangnaka [Manaflow](  []( "Twitter account") []( "LinkedIn profile") ### Company Launches [### Manaflow - Build and manage your dream AI operations team]( ### **🛎️ TL;DR** * [**Manaflow**]( empowers **operation managers** to automate workflows involving data analysis, API calls, and business actions. * You can **command Manaflow agents in English** to execute recurring tasks and manage them on a spreadsheet interface. * Think of us as an **AI-first Zapier alternative** but with natural language and spreadsheets. * **Frustrated with repetitive manual tasks** in your business? We’ll automate them all for you. [**Let’s chat**]( ### **🗂️ The Problem: Too much Excel, too much manual stuff, too much repetition** Are you tired of **juggling countless Excel files and manual workflows for your business**? Do these workflows prevent your business from scaling up? After **conducting over 250 calls**, we've discovered a recurring problem: **small-to-mid-sized businesses (SMBs)** rely heavily on folders of Excel files and third-party apps to **manually execute their day-to-day operations**, which is **time-consuming, error-prone, and a huge bottleneck to scaling**. For instance, in freight forwarders, operations include managing client communications, inventory oversight, and coordinating delivery schedules, most of which are done manually. These **manual processes** not only **decelerate business growth** but also **heighten the risk of mistakes**. Today’s operation managers are blocked by a lack of **technical knowledge,** **customization,** **and** **simplicity** in current workflow automation builders. ### **💡 Introducing MANAFLOW: Automate operations workflows with AI** We are building **Manaflow** specifically for **underdog SMBs to scale** like their tech-enabled, big corporation counterparts, transforming manual spreadsheet and software tasks into **automated end-to-end workflows**. The ideal way to execute workflows is not for a human to manually operate Excel sheets or interface with different software apps but for a human to **merely click a button**. **AI can own end-to-end technical workflows** and convert them into features, while **humans can oversee** them, update them as the business evolves, and tackle higher-level automations. Let’s see this in **action**. ### **📕 Building a Basic Workflow on Manaflow** We’ll begin by building a fun, basic workflow on Manaflow to **send payment reminders to clients** via email. [Manaflow Demo: Building Sending Payment Reminders Workflow]( Next, we’ll go into more specific examples of potential use cases for business customers. ### **📸 Manaflow Workflow #1: Watermark videos and send them to your clients via Gmail** One of our customers uses Manaflow to take in unpolished videos from Google Drive, process them, watermark them with their logo, and then email the final products to their clients — **all in one click of a button**. [Manaflow Demo: Watermarking Video Workflow]( This has **cut their** **20-hour weekly manual workflow down to 20 minutes**, and Manaflow has become a core part of their operations. ### **🛻 Manaflow Workflow #2: Find and email truckers for your shipment logistics** Another customer uses Manaflow to find and email local trucking companies based on the shipment origin and final destination. [Manaflow Demo: Finding Truckers for Logistics Workflow]( This has allowed them to gather rate quotations faster for their clients and has **reduced their manual operations by over 50%.** ### **🙌 One more thing: Collaborate with all stakeholders, including us, to build automations** There are countless workflow automations that you can build on Manaflow. We understand that people have many different ideas and love working in teams. So, we made **real-time collaboration** so all of you can build and automate workflows together on Manaflow. Partnering with us, you will also have **24/7 access to Manaflow’s engineering support staff.** ### 👥 **OUR TEAM** [**Austin Wang**]( Yale Physics, CS & Econ; Prev. SWE @ Google, NASA JPL, [Chess.com]( ML @ Datacy [**Lawrence Chen**]( Berkeley CS; Prev. Founding Eng @ Minion AI, Berkeley RISE Lab [**Wesley Tjangnaka**]( Stanford Math & CS; Prev. Stanford AI Research, ML @ Juniper Networks **🙏 Our Ask** **If you are or know any** **operation managers or businesses** **that run on spreadsheets**, **please let us know or schedule a time with us** [**here**]( Any intros or shoutouts would be super awesome and greatly appreciated. We’d really appreciate a like and subscribe on our [YouTube]( a follow on our [Twitter]( and [LinkedIn]( and some feedback on our [product]( Thank you! Footer ------ [Y Combinator]( "Y Combinator") ### Programs * [YC Program]( * [Startup School]( * [Work at a Startup]( * [Co-Founder Matching]( ### Company * [YC Blog]( * [Contact]( * [Press]( * [People]( * [Careers]( * [Privacy Policy]( * [Notice at Collection]( * [Security]( * [Terms of Use]( ### Resources * [Startup Directory]( * [Startup Library]( * [Investors]( * [SAFE]( * [Hacker News]( * [Launch YC]( * [YC Deals]( ### Make something people want. [Apply]( [Twitter]( © 2024 Y Combinator

Sepal AI

sepal-ai

https://www.sepalai.com

San Francisco, CA, USA

Most data requires domain knowledge that can be hard to source and curate, and publicly available benchmarks are contaminated or too general to be useful to actual product builders. Sepal AI is the data development platform that enables people to build useful datasets. We bring data generation tooling, synthetic data augmentation, rigorous quality control, and a network of over 20k PhD and industry experts into one platform so you can manage the production of high quality datasets.

The data development platform for LLM users and builders.

B2B

B2B

S24

Active

Early

https://www.ycombinator.com/companies/sepal-ai

https://yc-oss.github.io/api/batches/s24/sepal-ai.json

Title: Sepal AI: Building LLMs for Large Enterprises. | Y Combinator URL Source: Markdown Content: ### Building LLMs for Large Enterprises. Sepal AI builds Large Language Models for Enterprises through data development, finetuning, and inference. Our team comes from Turing, Vercel, McKinsey, and Bain. At Turing, we built the LLM training business and products to support over $120M revenue growth in 6 months for companies like Open AI, Google, and Anthropic. We learned that large, non-tech enterprises that we worked with, like PepsiCo, Bridgestone, and Volvo, don't have the data they need to train models to produce real value. Which means they’re not going to unlock the value from AI without a partner. We are targeting the 2400 largest non-software companies to build, continuously fine tune, and deploy their custom models. Sepal AI Founded:2024 Team Size:3 Location:San Francisco ### Active Founders ### Robi Lin, Founder Co-Founder @ Sepal AI Built the enterprise workflow products and fulfillment strategy at Turing.com. Scaled Turing’s LLM trainer business line from 50 to 800+ onboarded developers in 5 months for foundational LLM and enterprise customers. Previously was at Bain & Co. ### Kat Hu, Founder Cofounder @ Sepal AI Built Turing’s Foundational LLM trainer business GTM. Ran orgs of 500+ AI trainers & built corresponding operations for scale. Previously was at McKinsey. ### Fedor Paretsky, Founder I'm the co-founder and CTO at Sepal AI. Previously, I built platforms and infrastructure to bill users at Vercel while it went through hyper growth ($20M -\> $100M ARR). Before that, I worked on FP&A software at Mosaic and on platforms and infra at Newfront Insurance. ### Company Launches [### 🌱 Sepal AI - Confidently deploy your AI models]( **Tl;dr:** Sepal provides frontier data and tooling for advancing responsible AI development. --------------------------------------------------------------------------------------------- **\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_** ------------------------------------------------------------------------------------------------------------------------ [Sepal AI]( is on a mission to advance human knowledge and capabilities with the responsible development of artificial intelligence. **🧐 Responsibly advance human knowledge with AI? What does that mean?** ------------------------------------------------------------------------ We believe in a world where AI advances scientific research and empowers economic growth. To achieve that future, AI product & model builders need: 1. **Golden Datasets and Frontier Benchmarking:** To iteratively measure model performance on specific use cases. 2. **Training Data:** To improve model capabilities using fine-tuning and RLHF. 3. **Safety / Red-teaming:** To measure and forecast the safety of LLMs before putting them out in the wild. **\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_** ------------------------------------------------------------------------------------------------------------------------ ⚠️ **Okay, well why does it matter?** ------------------------------------- Frontier data for AI development is vital for safe deployment & scaling. However, developing this data is difficult. Most frontier data requires domain knowledge that can be hard to source and curate (e.g., finance, medical, physics, biology, etc.). Publicly available benchmarks (e.g., MMLU, GPQA, MATH, etc.) are contaminated and too general to be useful to actual product & model builders. **\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_** ------------------------------------------------------------------------------------------------------------------------ **🌱 How do we do this?** ------------------------- We’ve built Sepal AI - the data development platform that enables you to curate useful datasets. **The Platform:** We bring data generation tooling, human experts, synthetic data augmentation, and rigorous quality control into one platform so you can manage the production of high-quality datasets. **Our Expert Network:** We’ve built a network of 20k+ experts across STEM and professional services (think academic PhDs, business analysts, medical professionals, marketing and finance consultants) to support campaign design & data development. Sample engagements we’ve run: * **🧬 Cell and Molecular Biology Benchmark:** An original benchmark to evaluate complex reasoning across models. Produced by a team of PhD biologists from top institutions in the US. * **💼 Finance Q&A + SQL Eval:** A Golden Dataset to test the ability of an AI agent to query a database and produce human-expert-level answers to complex finance questions. * **📏 Uplift Trials & Human Baselining:** End to end support for conducting secure in-person evaluations on model performance. * _…. \[insert your custom use case next?\]_ **\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_** ------------------------------------------------------------------------------------------------------------------------ **🙏 Asks:** ------------ 1. If you are building an AI application and need to measure or improve your model, or 2. If you are a researcher at an AI lab building or evaluating models for new capabilities / risk areas, or 3. If you’re passionate about the development of AI, AI safety, or evals in general… Let’s [chat]( **\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_** ------------------------------------------------------------------------------------------------------------------------ **👪 Our team:** ---------------- Meet [Kat]( (on the left), [Robi]( (in the middle), [Fedor]( (on the right)! Robi and Kat previously built the technical LLM training business for Turing. Kat on the go-to-market & operations side. Robi on the product & fulfillment side. Fedor is a long-time close friend - he was an early engineer at Vercel & Newfront where he built out foundational infrastructure. Say hi: [[email protected]](mailto:[email protected]).

RetroFix AI

retrofix-ai

https://www.retrofix.ai/

San Francisco, CA, USA

RetroFix is a web platform that allows contractors to automatically apply for tax incentives and sustainability credits. Currently, contractors rarely apply for credits because discovering incentives, checking eligibility, and applying are all done manually via emails and phone calls with local utilities and/or government offices. RetroFix consolidates information on government incentives and allows building managers/contractors to apply in minutes, saving hundreds of thousands of dollars per building.

Turbo Tax for Building Rebates

Industrials

Industrials -> Climate

S24

Active

Early

https://www.ycombinator.com/companies/retrofix-ai

https://yc-oss.github.io/api/batches/s24/retrofix-ai.json

Title: RetroFix AI: Turbo Tax for Building Rebates | Y Combinator URL Source: Markdown Content: ### Turbo Tax for Building Rebates RetroFix is a web platform that allows contractors to automatically apply for tax incentives and sustainability credits. Currently, contractors rarely apply for credits because discovering incentives, checking eligibility, and applying are all done manually via emails and phone calls with local utilities and/or government offices. RetroFix consolidates information on government incentives and allows building managers/contractors to apply in minutes, saving hundreds of thousands of dollars per building. RetroFix AI Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Isaac Toscano, Founder Co-Founder/CEO at RetroFix AI. Before Retrofix I studied Materials Engineering at MIT, made financial models for billion dollar companies, did solar cell research at MIT's Nano lab and spend the weekends drilling Muay Thai. ### Daniel Portela, Co-Founder, CTO Co-Founder, CTO @ RetroFix AI | MIT CS | 🇵🇷 ### Company Launches [### 🏢 RetroFix AI - Turbo Tax for building rebates]( **tl;dr:** Building operations are responsible for around **1/4 of global energy CO2e** emissions. We want to make it extremely easy for every building in the world to decarbonize and decrease this number. Currently, we're working on making it extremely easy for buildings to secure government subsidies to become more energy efficient, starting with New York. [ Hi everyone, we’re [Isaac]( and [Daniel]( 👋 We met two years ago while studying at MIT, and now we are solving the problem of building decarbonization at [RetroFix AI]( **The Problem** 🌍 ------------------ Building operations are the 4th largest global contributor to carbon emissions: * **Globally**, they are responsible for around **26%** of all energy and process-related CO2e emissions * **In the US**, that number is approximately **29%** * **In dense cities** like New York, DC, Boston, and Philadelphia, they account for over **2/3** of all GHG emissions **Our Mission** 🚀 ------------------ We want to make it dead simple for every building in the world to decarbonize. We are committed to accelerating the adoption of clean energy technologies in buildings all around the world. We start by solving the largest barrier to decarbonization: money. Currently, we're working on making it extremely easy for buildings to secure government subsidies to become more energy efficient, starting with New York. 🏢 **Why Start with New York?** 🏙️ -------------------------------- With the passing of NYC’s Local Law 97, which penalizes building owners for exceeding emissions thresholds, around 12% of all buildings in NYC are starting to get fined this year for non-compliance. In 2024 alone, $20M in fines are estimated for these building owners. If no action is taken, the fines are expected to increase to over $85M per year by 2030, with over 70% of buildings being non-compliant. 💸 These fines will continue to rise as emissions thresholds become more stringent over time. Consequently, investments to improve critical building infrastructure have grown substantially, along with the need for subsidies. Consolidated Edison (the largest utility in New York) alone has deployed over $2B in subsidies over the last 4 years! 💡🏢 Contractors in NYC often spend hours researching the latest requirements and regulations to submit incentive applications, which are often not done correctly. As a result, a contractor may promise a certain amount of incentives to a customer but potentially deliver a completely different amount. This volatility in subsidy money is problematic for contractors who focus on delivering the best pricing and service to their clients. Because subsidy programs are often tied to government entities, contractors and building owners are interested in understanding how to “stack” incentives, but are deterred by the required reading to understand how to do so. 📚❗️ **The Solution ✅** ------------------ We start tackling the money problem by streamlining subsidy applications for energy-efficient equipment for buildings. Contractors no longer need to spend hours emailing or calling incentive programs, wondering when they'll hear back or if they did anything wrong! By using AI to understand regulatory constraints, we accurately identify the best incentives and provide a comprehensive list of required documentation. 📝 More time and money back in our contractors' pockets! We constantly update our software to ensure the most up-to-date information, so users remain worry-free. 😌 Contractors can rest assured they are submitting the right paperwork to maximize their clients’ incentives! Our goal is to accelerate the energy transition while saving owners and contractors hundreds of thousands of dollars through low-cost sustainability improvements. 🌿 **The Ask 🤲** -------------- We’re actively looking to onboard more building owners, building management companies, and HVAC professionals/engineers based in NYC! If you’re interested or know anyone we should connect with, please send us a message at [[email protected]](mailto:[email protected])! 📩 **The Team 🗣️** ----------------

Evolvere BioSciences

evolvere-biosciences

https://www.evolverebiosciences.com/

Oxford, England, United Kingdom

🦠🤖 We use our computational models to make next-generation antibiotics that outcompete bacterial evolution and precisely target pathogenic bacteria, without harming good microbes or human cells. ☠️ Current antibiotics stop working because bacteria evolve resistance to them. This makes drug-resistant bacteria a looming global health crisis - already killing more people than malaria and AIDS and it is getting exponentially worse 📈. 🧬 Our approach leverages co-evolutionary protein-protein interaction datasets combined with AI to forecast bacterial mutations and create ‘future-proof’ antibiotics, addressing antibiotic resistance before it develops. This changes the game for how frequently society will need to make new antibiotics and how long our new antibiotics will be able to treat patients 👩‍⚕️. We are a team of biochemists and evolutionary biologists who met at the University of Oxford.

Making Next-Generation Antibiotics that Outsmart Bacterial Evolution

Healthcare

Healthcare -> Therapeutics

S24

Active

Early

https://www.ycombinator.com/companies/evolvere-biosciences

https://yc-oss.github.io/api/batches/s24/evolvere-biosciences.json

Title: Evolvere BioSciences: Making Next-Generation Antibiotics that Outpace Bacterial Evolution | Y Combinator URL Source: Markdown Content: ### Making Next-Generation Antibiotics that Outpace Bacterial Evolution 🦠🤖 We use our computational models to make next-generation antibiotics that outcompete bacterial evolution and precisely target pathogenic bacteria, without harming good microbes or human cells. ☠️ Current antibiotics stop working because bacteria evolve resistance to them. This makes drug-resistant bacteria a looming global health crisis - already killing more people than malaria and AIDS and it is getting exponentially worse 📈. 🧬 Our approach leverages co-evolutionary protein-protein interaction datasets combined with AI to forecast bacterial mutations and create ‘future-proof’ antibiotics, addressing antibiotic resistance before it develops. This changes the game for how frequently society will need to make new antibiotics and how long our new antibiotics will be able to treat patients 👩‍⚕️. We are a team of biochemists and evolutionary biologists who met at the University of Oxford. Evolvere BioSciences Founded:2021 Team Size:3 Location:Oxford, United Kingdom ### Active Founders ### Adam Winnifrith, Founder Evolvere (S24), CEO & Co-Founder, MBioChem Oxford, I work at the intersection of bioengineering, protein AI, and automation. E.S.B ### Piotr Jedryszek, Founder S24, Evolvere Biosciences, CTO and Co-founder, background in Bio and CompBio ### Weronika Slesak, Founder ex-Oxford biologist using evolution as an engineering tool. ### Company Launches [### 🦠💊 Evolvere Biosciences – Making next-generation antibiotics that outsmart bacterial evolution]( Hi everyone! – We’re [Piotr]( [Weronika]( and [Adam]( a team of biochemists and evolutionary biologists from the University of Oxford on a mission to make the next-generation of antibiotics. Current antibiotics stop working because bacteria evolve resistance to them. Our approach leverages co-evolutionary **protein-protein interaction datasets** combined with **AI** to forecast bacterial mutations and create **‘future-proof’ antibiotics**, addressing antibiotic resistance before it develops. This changes the game for how frequently we’ll need to make new antibiotics and how long our new antibiotics will be able to treat patients. Let's get into more detail: ❌ What’s the problem? --------------------- Antibiotic resistance is a looming global health crisis: * ☠ Already killing more people than Malaria and HIV * 📈 Getting exponentially worse because of bacterial resistance * 💸 $100 trillion economic burden undermining modern medicine ✅ We solve this by making future-proof antibiotics that: -------------------------------------------------------- 1. 🏃‍♂ Stay ahead of bacterial mutations to prevent resistance 2. 🎯 Precisely target harmful bacteria without disrupting beneficial microbes 3. 💵 Overcome the economic challenges of antibiotic development **Our Science** --------------- Traditional trial-and-error discovery cannot compete with bacteria's ability to mutate and acquire resistance genes. Our **evolutionary datasets** and **AI** will allow us to stay one step ahead of bacteria. We don’t react, we anticipate: You might wonder whether bacteria would eventually be able to mutate in other ways around our antibiotics. Well, yes, they could, but our approach forces all the escape mutations to be extremely costly. In fact, so costly that the bacteria wouldn’t survive. How? Our experiments are like running a battle simulation hundreds of times to find enemies’ weak points. This means that we can create detailed maps of the co-evolutionary landscapes of bacteria and our antibiotics so that we can ultimately engineer medicines with a low propensity for resistance emergence. [Watch how that works:]( We then engineer our antibiotics for **stability and safety** inside the human body using a suite of **protein AI models** (both diffusion and language model-based). This engineering means our antibiotics 1) only target pathogenic bacteria and not human cells or microbiomes and 2) have the potential to be given as a single dose – reducing the amount of monitoring that doctors have to do on patients. This is in contrast to current antibiotics, which can have human cell toxicity, disrupt microbiomes, and have to be dosed every few hours. * * * 👩‍⚕️ Why doctors are excited ----------------------------- Our blueprints have the potential for: * Low drug-drug interactions * Low risk of _C. difficile_ infection * Low dosing regimes * Low risk of resistance emergence * Low side effect profile 📊 Evolvere Bio Factfile ------------------------ * 🧪 We have already synthesized molecules that specifically kill bacteria in **physiological conditions**. These molecules are specific to only their target bacteria. * 🧬 Our R&D generates valuable **data** on protein co-evolution. * 🤖 We build **AI models** to predict and prevent resistance that arises from **protein co-evolution**, with potential applications in other therapeutic areas. Team ---- * 🧑‍🔬 **Adam Winnifrith** - Oxford biochemist who developed new biochemical assays based on advanced statistical concepts and published work on the advancements in generative AI in protein design. * 🧑‍🔬 **Piotr Jedryszek** – Oxford computational biologist studying the evolution of bacteria using deep learning techniques. His past work included molecular dynamics simulations, nanopore engineering, and biofuels. * 👩‍🔬 **Weronika Ślesak** – evolutionary biologist who worked at renowned microbiology laboratories (at the University of Oxford and Institut Pasteur) on high-throughput experimental evolution, antibiotic resistance genes and evolutionary trade-offs. * 👨‍💼 **Oliver Waterhouse** - serial biotech entrepreneur who sold his previous Oxford-based company Base Genomics for $410 million. Our Asks -------- Are you as excited as we are about making future-proof antibiotics? * Upvote/share this post! * Reach out to us at [[email protected]](mailto:[email protected])

Conductor Quantum

conductor-quantum

https://www.conductorquantum.com

San Francisco, CA, USA

Quantum computers will allow humanity to simulate nature at the quantum level, enabling the acceleration of drug discovery and the development of new materials. Currently, quantum engineers spend days, sometimes weeks, manually getting their silicon chips to operational conditions to realize two qubits. A qubit is the information-carrying unit of a quantum computer, analogous to a bit in a classical computer. We need billions of qubits to make a useful quantum computer. Therefore, automation software will be vital to realizing this goal. Conductor Quantum will develop AI software to remove the human from the loop, enable the scaling of silicon quantum technology and build a silicon-based quantum computer.

Quantum computers on silicon chips using AI.

Industrials

Industrials

S24

Active

Early

https://www.ycombinator.com/companies/conductor-quantum

https://yc-oss.github.io/api/batches/s24/conductor-quantum.json

Title: Conductor Quantum: Quantum computers on silicon chips using AI. | Y Combinator URL Source: Markdown Content: ### Quantum computers on silicon chips using AI. Quantum computers will allow humanity to understand the world at its most fundamental level, enabling the acceleration of drug discovery and the development of new materials. Currently, quantum engineers spend days, sometimes weeks, manually getting their silicon chips to operational conditions to realize two qubits. A qubit is the information-carrying unit of a quantum computer, analogous to a bit in a classical computer. We need billions of qubits to make a useful quantum computer. Therefore, automation software will be vital to realizing this goal. Conductor Quantum will develop AI software to remove the human from the loop, enable the scaling of silicon quantum technology and build a silicon-based quantum computer. Conductor Quantum Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Brandon Severin, Founder Co-founder of Conductor Quantum, where we build quantum computers on silicon chips using AI. During my PhD at Oxford, I worked with 4 institutions across the globe (IST Austria, Basel, UNSW, Diraq), developing algorithms for semiconductor quantum device control and published 4 papers on the topic including one in Nature. ### Joel Pendleton, Founder Co-founder of Conductor Quantum, building quantum computers on silicon chips with AI. I’ve worked at many deep tech startups and research labs, exploring various quantum computing technologies — from carbon nanotubes to superconducting transmon qubits. I left my PhD at Oxford to start Conductor Quantum. ### Company Launches [### Conductor Quantum - Quantum computers on silicon chips using AI]( **Tl;dr:** * [Conductor Quantum]( is using **AI software** to **create qubits** in **semiconductor chips** to **build quantum computers** at scale. * The two founders met during their **PhDs at Oxford**. They have experience at 5 quantum companies and institutions, plus 4 publications between them. * Conductor Quantum is launching **the first publicly available API** to **classify quantum transport features in semiconductors**. **🧑‍🔬🧑‍💻The Team** ---------------------- Hi - We are Brandon and Joel! We are building Conductor Quantum. CEO, [Brandon]( – Brandon developed **_AI for Quantum Computing in Silicon_** during his **PhD** at **Oxford**. Over the past four years, Brandon has worked with four top quantum research institutions on **software for quantum device control** and has **four publications**, including one in **Nature**. CTO, [Joel]( – Joel has shipped software at **four deep tech companies** and has worked with a series of quantum devices ranging from **carbon nanotubes to superconducting circuits**. Joel studied **Physics at UCL** and **dropped out of his PhD** at **Oxford** to start Conductor Quantum. During Joel’s PhD at **Oxford**, he launched [Feynman.ai]( an AI science research assistant— with Brandon, which they bootstrapped to **300 signups within a month** of launch. **😯 The Problem** ------------------ * **Quantum computers** will **revolutionize drug discovery** and **material development** as we will, for the first time, be able to accurately **simulate nature at its most fundamental level**. * Leveraging the **trillion-dollar semiconductor** industry, silicon chips offer a **scalable architecture** for building quantum computers. → Currently, quantum engineers **spend days, sometimes weeks,** manually configuring silicon chips to **create a single quantum bit (qubit)**, analogous to a bit in a classical computer. **→ We need millions, if not billions, of qubits** to make a **useful quantum computer.** Automation software to create qubits at scale is vital to realize this goal. **✅ Our Solution - AI software to create qubits** ------------------------------------------------- * We are building **leading AI software** that creates qubits by **learning and understanding the principles of quantum transport** in semiconductor chips. * Our AI software will **replace the human** in the loop, **unlocking rapid qubit generation** and **fabrication feedback iteration cycles**. * Automatic qubit creation and control are the foundation of **the first quantum operating system** on **semiconductor chips**. [Our AI Playground]( **🤝 How You Can Help** ----------------------- * Sign up for our [**waitlist**]( (Announcements soon!) * Quantum engineers who want to get back to building and save themselves 100s of hours per qubit created - Email us at [[email protected]](mailto:[email protected]) * We’d love intros to: * Chip designers/verifiers/manufacturers * Anyone who wants to **bring chip manufacturing back to the USA** ### Hear from the founders #### What is your long-term vision? If you truly succeed, what will be different about the world? Everyone will have access to a quantum computer from their desk. ### Selected answers from Conductor Quantum's original YC application for the S24 Batch #### How long have each of you been working on this? How much of that has been full-time? Please explain. We’ve been working on the problem of quantum device control individually for the past 4 years as part of our degrees and work experience. Brandon (4 years full-time), Joel (2 years full-time, 2 years part-time).

Moonglow

moonglow

https://moonglow.ai

San Francisco, CA, USA

Moonglow connects local Jupyter notebooks to remote cloud compute. Machine learning researchers and data scientists use us to scale up their experiments without having to do DevOps. Previously, Leila was a software engineer at Jane Street. She led the build-out of its equities clearinghouse connectivity infrastructure and was the technical lead of its front-office SRE team. Trevor was part of Hazy Research Lab at Stanford, and published machine learning research at ACL and ICLR. With the explosive growth of deep learning, there are over 1 million ML researchers who do computationally intensive experiments every day. They use Jupyter notebooks to do so, but every time they want to try out a new idea, they need to get the notebooks running on cloud machines. This process, repeated hundreds of times a month, is time-consuming and error-prone. Moonglow reliably brings the time-to-experiment down from 5 minutes to 20 seconds. Just as Vercel and Replit abstracted away the lower levels of the computing stack for web developers and programmers, we do the same for ML researchers. https://moonglow.ai

Connecting local Jupyter notebooks to remote cloud compute.

B2B

B2B -> Engineering, Product and Design

S24

Active

Early

https://www.ycombinator.com/companies/moonglow

https://yc-oss.github.io/api/batches/s24/moonglow.json

Title: Moonglow: Connecting local Jupyter notebooks to remote cloud compute. | Y Combinator URL Source: Markdown Content: ### Connecting local Jupyter notebooks to remote cloud compute. Moonglow connects local Jupyter notebooks to remote cloud compute. Machine learning researchers and data scientists use us to scale up their experiments without having to do DevOps. Previously, Leila was a software engineer at Jane Street. She led the build-out of its equities clearinghouse connectivity infrastructure and was the technical lead of its front-office SRE team. Trevor was part of Hazy Research Lab at Stanford, and published machine learning research at ACL and ICLR. With the explosive growth of deep learning, there are over 1 million ML researchers who do computationally intensive experiments every day. They use Jupyter notebooks to do so, but every time they want to try out a new idea, they need to get the notebooks running on cloud machines. This process, repeated hundreds of times a month, is time-consuming and error-prone. Moonglow reliably brings the time-to-experiment down from 5 minutes to 20 seconds. Just as Vercel and Replit abstracted away the lower levels of the computing stack for web developers and programmers, we do the same for ML researchers. Moonglow Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Leila Clark, Co-Founder I'm currently building Moonglow, which connects Jupyter notebooks to cloud compute. Before this, I was a software engineer at Jane Street, where I led the build-out of its equities clearinghouse connectivity infrastructure and was the technical lead of its front-office SRE team. I graduated from Princeton with highest honors in Computer Science. ### Trevor Chow, Co-Founder Trevor is the co-founder of Moonglow. Previously, he graduated with a BS in Mathematics from Stanford, where he was part of Hazy Research Lab and published machine learning research at ACL and ICLR. He also traded index options and optimized low latency execution strategy at Optiver. ### Company Launches [### Moonglow - Colab on your cloud]( Hey everyone, we’re Trevor and Leila from [Moonglow]( ❌ Problem: moving experiments to cloud GPUs sucks ------------------------------------------------- When you’re doing machine learning research, it’s important to try out new ideas quickly. Jupyter notebooks make that easy. But what happens when your local computer isn’t enough? Your workflow probably looks like this: 1. Go to your cluster or cloud provider 2. Pick the right configs and spin up a node 3. SSH into the node 4. Install all the required packages 5. Pull your code from GitHub All of this is before you’ve even run a single cell in your notebook! And if you want to share your work or come back to it later, either you need to keep your GPU running (expensive) or go through this entire process again (slow). 🎉 Solution: Bring your own compute to Jupyter ---------------------------------------------- Moonglow connects your local Jupyter notebooks to your cloud compute provider. With a click of a button, you can switch runtimes and scale up your experiments to the GPUs you need. We handle all of the messy DevOps under the hood, and since your notebook lives in your local IDE, you can easily come back to it and get it running in seconds! We currently support connecting notebooks in VS Code / Cursor to Runpod instances, and we’re expanding this to other providers soon e.g. AWS, GCP, Azure, Lambda Labs etc. 👀 Team ------- [Trevor]( used to do ML research at Stanford, while [Leila]( was a software engineer working on high-performance infrastructure at Jane Street. We started Moonglow because we’ve seen how janky and unintuitive the current tooling is for ML research, and how that is bottlenecking the pace at which researchers can validate their results at scale. 🙏 Asks ------- * [**Try out**]( Moonglow or [book a time]( to get set up. * Let us know which **cloud providers** we should support next! * Connect us to **ML researchers** you know. We’re excited to hear from you, either at [[email protected]](mailto:[email protected]) or on Linkedin ([Trevor]( [Leila](

Hamming AI

hamming-ai

https://hamming.ai/

San Francisco, CA, USA

Humans make billions of calls/day. We think a majority of these will be handled by AI built by thousands of companies tackling every single vertical. Making these AI voice agents reliable is hard. A small change in prompts, function call definitions, or model providers can cause large changes in LLM outputs. Hamming automates testing for AI voice agents. Our voice agents call your voice agent. An AI drive-through startup uses Hamming to simulate thousands of simultaneous phone calls to achieve 99.99% agent order accuracy. We have a proven track record of helping enterprises win with AI. Sumanyu (CEO) previously helped Citizen (safety app) grow its users by 4X and grew an AI-powered sales program to 100s of millions in revenue/year at Tesla. Marius (CTO) previously ran data infrastructure @ Anduril and was a founding engineer @ Spell (MLOps startup acquired by Reddit).

Automated testing for AI voice agents

B2B

B2B -> Engineering, Product and Design

S24

Active

Early

https://www.ycombinator.com/companies/hamming-ai

https://yc-oss.github.io/api/batches/s24/hamming-ai.json

Title: Hamming AI: Automated testing for AI voice agents | Y Combinator URL Source: Markdown Content: Hamming AI: Automated testing for AI voice agents | Y Combinator =============== []( "Y Combinator") [About]( [What Happens at YC?]( Interview Guide]( Blog]( [Companies]( [Startup Directory]( Companies]( Directory]( YC]( [Startup Jobs]( [All Jobs]( Engineering]( Operations]( Marketing]( Sales]( Internship Program 2024]( Job Guide]( Coaching]( Startup Jobs Blog]( [Find a Co-Founder]( [Library]( [SAFE]( [Resources]( [Startup School]( Investors]( News]( Open main menu Apply for **W2025** batch.[Apply]( "Apply for W2025 batch.") [Home]( AI Hamming AI ========== Automated testing for AI voice agents [S24]( Active [developer-tools]( Francisco]( * * * [Company]( [Jobs](  [ * * * ### Automated testing for AI voice agents Humans make billions of calls/day. We think a majority of these will be handled by AI built by thousands of companies tackling every single vertical. Making these AI voice agents reliable is hard. A small change in prompts, function call definitions, or model providers can cause large changes in LLM outputs. Hamming automates testing for AI voice agents. Our voice agents call your voice agent. An AI drive-through startup uses Hamming to simulate thousands of simultaneous phone calls to achieve 99.99% agent order accuracy. We have a proven track record of helping enterprises win with AI. Sumanyu (CEO) previously helped Citizen (safety app) grow its users by 4X and grew an AI-powered sales program to 100s of millions in revenue/year at Tesla. Marius (CTO) previously ran data infrastructure @ Anduril and was a founding engineer @ Spell (MLOps startup acquired by Reddit). Hamming AI Founded:2024 Team Size:2 Location:San Francisco Group Partner:[Gustaf Alstromer](  []( "LinkedIn profile") []( "Twitter account") []( "Crunchbase profile") []( "Github profile") ### Active Founders ### Sumanyu Sharma, Founder Sumanyu is the Co-Founder & CEO @ Hamming. Previously helped Citizen grow its MAU by 4X and helped bootstrap revenue from 0 to millions in ARR in under 6 months. Before that, grew an AI-powered sales program @ Tesla to 100s of millions in revenue/year as a Senior Staff Data Scientist. Published a first-author paper in AI during undergrad. BASc from UWaterloo w/ dean's list. Sumanyu Sharma [Hamming AI](  []( "Twitter account") []( "LinkedIn profile") ### Marius Buleandra, Co-Founder & CTO Marius is the Co-Founder & CTO @Hamming. Previously Eng Manager for Data Infrastructure @Anduril. Founding engineer @Spell (ML Observability & Infra startup acquired by Reddit). Worked on payments @Square and Windows Kernel Virtualization @Microsoft. Marius Buleandra [Hamming AI](  []( "Twitter account") []( "LinkedIn profile") ### Company Launches [### 🕵️ Hamming - Automated testing for voice agents]( 👋 [@Sumanyu Sharma]( and [@Marius Buleandra]( from [@Hamming AI]( **TLDR:** Are you testing your voice agents by hand? We're launching [Voice Simulations]( to automatically test your voice agents and flag quality issues in development and production. 🌟 [Click here to try our free Voice Simulations Demo]( 🌟 Problem: Making voice agents reliable feels like whack-a-mole ------------------------------------------------------------- Here's the workflow most teams follow: 1. **Call** your voice agent by hand and find bugs. Slow and ad-hoc. 2. **Tweak** your voice agents by adding new tools and changing the prompts or models to fix the bugs. 3. **Call again** to see if the changes worked. 4. **Detect** regressions when users complain of things breaking in production. 5. **Repeat** steps 1 to 4 until you get tired. **Calling your voice agent & finding bugs** is the slowest & most painful part of the feedback loop. This is what we automate. Our take: Character AI for voice testing ---------------------------------------- We create hundreds of characters that simulate how **real users interact with your voice agents** in real life. For every call, we measure whether our character successfully accomplishes the task (e.g., ordering a vegan burger, canceling next week’s appointment, etc.). Our approach is 100x faster, cheaper, and more thorough than manual testing. ### Flag errors & Tag calls in production You can log all call transcripts and traces within Hamming. We actively **tag your production calls** in real-time, and flag cases the team needs to double-click on. This helps engineering teams quickly prioritize cases they need to fix. **Example tags**: human detects that the bot is an AI, a follow-up call is needed, the user requested an urgent appointment, etc. ### Test new changes quickly **Simulation-driven development** Let’s imagine you’re building an agent called ‘YC Founder’; we can spin up 100s of VC agents who will try to distract you. You can edit the prompts or models and re-run the simulation to make sure you made progress. Want to see how you would handle a persistent investor? Try our ‘VC trying to distract founders’ free [demo here]( **Easily create new characters from call transcripts** When customers complain about a bad call, you can locate the call transcript and create a new character in one click. Make a change to your prompt, and then run the simulations to ensure you addressed the bad call. Meet the team ------------- [Sumanyu]( previously helped Citizen (safety app; backed by Founders Fund, Sequoia, 8VC) grow its users by 4X and grew an AI-powered sales program to $100s of millions in revenue/year at Tesla. [Marius]( previously ran data infrastructure @ Anduril, drove user growth at Citizen with Sumanyu and was a founding engineer @ Spell (MLOps startup acquired by Reddit). Summary ======= We previously launched [Prompt Optimizer]( and [AI Experimentation Tools]( to automate prompt engineering and make RAG pipelines more robust. In this launch, we show how you can test your voice agents quickly. Our offer ========= **Personalized characters + 100 free calls.** Struggling to make your voice agents reliable? We’ll create personalized characters and call + stress test your system ~100 times for free. Book time with us [here.]( Questions? Email us [here](mailto:[email protected]) or chat with us [here]( ### Other Company Launches ### 🚀 Hamming - Make your RAG & AI agents reliable The only end-to-end AI development platform you need: prompt management, evals, observability [Read Launch ›]( ### 🚀 Hamming - Let AI optimize your prompts (free for 7 days) Automate 90% of manual prompt engineering using our self-improving prompt optimizer. [Read Launch ›]( Footer ------ [Y Combinator]( "Y Combinator") ### Programs * [YC Program]( * [Startup School]( * [Work at a Startup]( * [Co-Founder Matching]( ### Company * [YC Blog]( * [Contact]( * [Press]( * [People]( * [Careers]( * [Privacy Policy]( * [Notice at Collection]( * [Security]( * [Terms of Use]( ### Resources * [Startup Directory]( * [Startup Library]( * [Investors]( * [SAFE]( * [Hacker News]( * [Launch YC]( * [YC Deals]( ### Make something people want. [Apply]( [Twitter]( © 2024 Y Combinator

Blast

blast

https://blastsec.com

San Francisco, CA, USA

Blast helps large enterprises build safe and reliable LLM apps. We provide a platform to help enterprises rigorously evaluate and turn their generative AI prototypes into reliable apps that can be confidently deployed at scale.

Building safe and compliant LLM apps/agents for enterprises

B2B

B2B -> Security

S24

Active

Early

https://www.ycombinator.com/companies/blast

https://yc-oss.github.io/api/batches/s24/blast.json

Title: Blast: Building safe and compliant LLM apps/agents for enterprises | Y Combinator URL Source: Markdown Content: ### Building safe and compliant LLM apps/agents for enterprises Blast helps large enterprises build safe and reliable LLM apps. We provide a platform to help enterprises rigorously evaluate and turn their generative AI prototypes into reliable apps that can be confidently deployed at scale. Blast Founded:2024 Team Size:2 Location:San Francisco ### Active Founders ### Arnav Joshi, Founder CEO at Blast. I was recently a BS/MS student in CS at Stanford. Previously, I worked on autonomous driving research at NVIDIA. I also interned at Amazon, where I built a computer vision tool. ### Daniel Zamoshchin, Founder CTO at Blast. I studied computer science with a focus in security at Stanford. I worked on security research with Prof. Dan Boneh and data privacy research with Prof. Matei Zaharia (Databricks). In high school, I built an app with 2M+ downloads. ### Company Launches [### 💥 Blast - Deploy AI safely and reliably at scale]( **TL;DR:** [Blast]( provides a platform to help enterprises turn their generative AI prototypes into reliable apps that can be confidently deployed at scale. — Hi everyone, we're Daniel and Arnav, and we're building Blast. **Problem: The mass adoption of LLMs across Fortune 500s is blocked by safety and reliability concerns.** Every large enterprise is trying to deploy their LLM-based apps/agent POCs at scale. Unfortunately, prohibited content, hallucinations, and other failures leave enterprises open to brand damage and legal liability. For example, Air Canada was forced to honor an out-of-policy bereavement discount that its chatbot had offered to a grieving passenger. Legacy enterprises in particular lack the talent and the tooling to solve these issues. **Solution:** We’re helping large enterprises build more reliable LLM apps. 1. **Evaluation platform:** Our red teaming tools enable developers to probe their end-to-end system with multi-turn conversations aimed at uncovering content/policy violations. 2. **Governance models:** We help enterprises detect, fix, and log failures in production. We started two months ago, and we are piloting with a Fortune 50 company. **The Team:** 👋 Hey, it’s [Daniel]( and [Arnav]( We met ten years ago in middle school and have been close friends ever since. We went to high schools in different states but reconnected at Stanford, where we were freshman year roommates. Daniel previously worked on AI security research at Stanford. Arnav previously worked on autonomous driving research at NVIDIA and in macro at Bridgewater. **Our Asks:** * We’d love intros to anyone working in security or AI governance at large enterprises. Email us at [[email protected]](mailto:[email protected]) * If you’re interested in adversarially testing your AI apps/agents or evaluating your guardrails, we’d love to talk! Please email us above!

Saturn

saturn

https://www.SaturnOS.com

London, England, United Kingdom

At Saturn, our mission is to make financial peace of mind more accessible by building the best operating system for wealth managers. Wealth and Asset Management is one of the world’s highest revenue-generating industries, yet it remains painfully inefficient. The sector relies heavily on disjointed legacy systems, manual human tasks, and a complex web of intermediaries, all of which increase costs. Saturn is an AI-powered operating system for wealth management built to automate investment research, streamline operations, and ensure compliance. Saturn creates the best version of truth, empowering investment advisors to work more efficiently, navigate complex regulatory landscapes, and deliver personalised, high-value services to their clients. Today, Saturn supports over 200 firms globally that manage more than £35bn in assets under management (AUM) and over 200,000 clients. As the industry grows, it faces massive operational challenges like bad tech, advisor retirements, regulatory pressure, an increased cost base, and transformational opportunities like great intergenerational wealth transfer and changing consumer needs. Saturn aims to be at the forefront of this transformation, driving the future of wealth management.

Backoffice and compliance automation for wealth managers

Fintech

Fintech

S24

Active

Early

https://www.ycombinator.com/companies/saturn

https://yc-oss.github.io/api/batches/s24/saturn.json

Title: Saturn: Compliance and back office workflows for Wealth Managers. | Y Combinator URL Source: Markdown Content: ### Compliance and back office workflows for Wealth Managers. At Saturn, our mission is to make financial peace of mind more accessible by building the best operating system for wealth managers. Wealth and Asset Management is one of the world’s highest revenue-generating industries, yet it remains painfully inefficient. The sector relies heavily on disjointed legacy systems, manual human tasks, and a complex web of intermediaries, all of which increase costs. Saturn is an AI-powered operating system for wealth management built to automate investment research, streamline operations, and ensure compliance. Saturn creates the best version of truth, empowering investment advisors to work more efficiently, navigate complex regulatory landscapes, and deliver personalised, high-value services to their clients. Today, Saturn supports over 200 firms globally that manage more than £35bn in assets under management (AUM) and over 200,000 clients. As the industry grows, it faces massive operational challenges like bad tech, advisor retirements, regulatory pressure, an increased cost base, and transformational opportunities like great intergenerational wealth transfer and changing consumer needs. Saturn aims to be at the forefront of this transformation, driving the future of wealth management. Saturn Founded:2023 Team Size:8 Location:London, United Kingdom ### Active Founders ### Amal Jolly, Founder/CEO Financial peace of mind enables every human to achieve more; we are on a mission to make it happen! Experienced in Product/GTM in Banking, Insurance, Wealth/Asset Management and Capital Markets technologies. ### Michael Ettlinger, Founder CTO & Co-Founder at Saturn AI Launched AI products used by 20M+ users across various industries. Background in AI research, focused on AI Agents since 2017. ### Rohit Vaish, Founder President & Co-founder at Saturn. Rohit has expertise in building investment portfolios for Wealth Management firms having spent over 5 years at BNY Mellon. Alongside that, he also led firm wide compliance initiatives across the UK fund management company to assess value for money and client suitability. After leaving financial services, Rohit led Product and Operations teams in a variety of startups before founding Saturn. ### Company Launches [### 🪐 Saturn: Backoffice and Compliance automation for Wealth Managers]( **tldr:** [Saturn]( automates investment research, compliance, and back-office workflows for wealth managers, significantly reducing operational costs and freeing up time to focus on client engagement and growth. — * * * * * * * * * Hey everyone, we're [Michael]( [Rohit]( and [Amal]( and we are on a mission to make wealth building more accessible for everyone by making wealth managers more efficient and scalable. 🧨 The Problem: Servicing a client with compliance and back-office tasks is a time drain ------------------------------------------------------------------------ Wealth managers spend an inordinate amount of time and resources on manual compliance and administrative tasks. This inefficiency leads to higher operational costs and limits their ability to serve more clients effectively. Take our client, Jason, as an example. Jason runs a mid-sized advisory firm in Norwich, UK, with over 15,000 clients and spends over 15 hours per client per year on compliance paperwork and back-office management all for doing one check-in a year. These manual processes drain his time and increase operational costs. This time and money could be better spent on growing his client base and enhancing client services. 🎉 The Solution: Automated AI-powered wealth management operating system built for scale and personalisation ------------------------------------------------------------------------------------------- Saturn helps wealth managers like Jason automate their compliance and back-office tasks, allowing them to focus more on client engagement and growth. Our platform uses AI to streamline processes and ensure real-time compliance, including: * Personalised investment research * Faster client onboarding * Seamless Compliance reporting * Client growth & analytics **Using Saturn, wealth managers:** ✔️ Save over 50% of the time typically spent on compliance and admin tasks ✔️ Increase operational efficiency, allowing them to offer more services and serve more clients ✔️ Improve personalisation, retention and net revenue from their client 💎 Opportunity: Driving the Future of Wealth Management --------------------------------------- Wealth management firms, like Jason's, waste countless hours and resources on manual compliance and operations—a massive inefficiency across the industry. Saturn is uniquely positioned to eliminate this burden through complete AI powered backoffice automation. Imagine a world where wealth managers can seamlessly guide clients through their entire financial lifecycle, making this service accessible to many more people. This is the future Saturn aims to create. > Big opportunities like the wealth transfer, growing consumer needs amongst wider demographics requires a more digital and robust system to serve them. > > The current systems simply cannot scale to serve the need, they are plagued by operational challenges added by bad tech, increasing regulatory pressure, and cost base. > > This industry is in a dire need of a transformation > > **Saturn aims to be at the forefront of this transformation, driving the future of wealth management.** **✅ Progress so far** 1. Grown to serve 200 firms globally 2. Cashflow positive since month 2 in the market, live since Feb. 3. Awesome team of 8 people. 👋 **Ask: How you can help** 1. Connect us to wealth managers or financial advisors in your network. [Intro us.]( 2. Know any killer engineer or product visionaries with experience in Wealth Tech and Financial Services. [Intro us.](mailto:[email protected]) ### Hear from the founders #### What is your long-term vision? If you truly succeed, what will be different about the world? Managing wealth and access to financial services will be a standard for all.

Ontra Mobility

ontra-mobility

https://www.ontramobility.com

New York, NY, USA; Remote

Ontra Mobility is founded by two former Googler engineers with PhDs in operations research. Ontra helps cities and other transit agencies increase ridership through data-driven planning and real-time optimization. The founders developed the software that powered MARTA Reach in Atlanta, GA and CAT SMART in Savannah, GA.

Platform for planning and operating multimodal transit systems

Government

Government

S24

Active

Early

https://www.ycombinator.com/companies/ontra-mobility

https://yc-oss.github.io/api/batches/s24/ontra-mobility.json

Title: Ontra Mobility: Platform for planning and operating multimodal transit systems | Y Combinator URL Source: Markdown Content: ### Platform for planning and operating multimodal transit systems Ontra Mobility is founded by two former Googler engineers with PhDs in operations research. Ontra helps cities and other transit agencies increase ridership through data-driven planning and real-time optimization. The founders developed the software that powered MARTA Reach in Atlanta, GA and CAT SMART in Savannah, GA. Ontra Mobility Founded:2023 Team Size:2 Location:New York ### Active Founders ### Anthony Trasatti, Founder Founder at Ontra Mobility ### Connor Riley, Founder Founder Ontra Mobility. Xoogler, PhD from @GeorgiaTechISyE. Formerly @UMIOE and @UConn ### Company Launches [### Ontra Mobility - Increasing ridership for transit agencies]( **tldr:** [Ontra]( helps cities and transit agencies increase ridership through data-driven planning and real-time optimization. **🚀 Introduction** ------------------- Hey everyone! We’re Anthony and Connor, aka the [Ontra Mobility]( team, and we’re revolutionizing how people move. **❗ The Problem** ----------------- Transit agencies in the U.S. are facing massive budget shortfalls (up to $3B). Agencies need to adapt their transit systems more quickly to meet the demands of our evolving cities. Work from home led to a large drop in ridership, which is problematic for train and bus routes, which have high fixed costs. Cutting lines or reducing frequency further impacts ridership - transit agencies need a different solution. **💡 The Solution** ------------------- Ontra helps cities and other transit agencies increase ridership through data-driven planning and real-time optimization. Our platform optimizes cost and rider experience by designing high-frequency bus routes and ridesharing zones, where our algorithms adapt routes to demand in real time. **⚙️ How It Works** ------------------- Ontra analyzes information about where people live, how they've traveled in the past, and the existing transit system to figure out the best way to redesign the transit network. Once the improvements are made, Ontra Mobility plans rider journeys in real-time with the best combination of transit options available, whether it’s a train, bus, bike, or agency-operated on-demand. Our system can handle thousands of ridesharing requests per minute – efficiently assigning riders to optimal routes. [ **🚌 What’s New** ----------------- * Our transit network design application generates bus routes and micro-transit zones * Seamless multimodal journeys – our algorithm synchronizes transfers between on-demand and fixed-route **🎓 The Team** --------------- Anthony and Connor are both former Google engineers with PhDs in [operations]( [research]( Prior to joining YC, they worked on [MARTA Reach]( in Atlanta, GA, and [CAT Smart]( in Savannah, GA. **🙏 Ask: How you can help** ---------------------------- * Share this post and spread the word! * Connect us to private- or public- **transit agency employees** or **board members**, **politicians,** and **transit advocates**. [[email protected]](mailto:[email protected]) Quick blurb to copy and paste: [Ontra Mobility]( is founded by two former Google engineers with PhDs in [operations]( [research]( Ontra helps cities and other transit agencies increase ridership through data-driven planning and real-time optimization. The founders developed the software that powered [MARTA Reach]( in Atlanta, GA and [CAT Smart]( in Savannah, GA. ### Hear from the founders #### How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?) Connor and Anthony met at Georgia Tech where they graduated with PhDs in operations research with a focus on optimization for public transit systems. Their publications have been cited over 150 times, appearing in IJACI and Transportation Science. Together, they have over a decade of experience in optimization of transit systems. While at Georgia Tech, they developed and managed full-stack deployments of MARTA Reach ([ a public on-demand ride-sharing service in Atlanta, GA, and the SMART Transit pilot ([ in Savannah, GA. ### YC S24 Application Video