{"id":127854,"title":"Epsilab - RL environments for enterprise work","tagline":"High-fidelity RL environments from real-life data and domain expertise","body":"**TLDR: Epsilab is an applied AI research lab building high-fidelity RL environments for long-horizon enterprise work. We partner with the world’s top professionals and work directly with frontier labs to improve model agentic capabilities in the real world.**\n\nWe believe in human judgement and taste, and have seen that current frontier models have a long way to go when deployed in the real world. The widespread diffusion of AI depends on distilling the knowledge and expertise of the top performers in every field.\n\n#### The problem\n\nToday, low data quality is a pervasive, industry-wide problem. Widely used benchmarks like [TerminalBench 4.0](https://preferencemodel.com/blog/terminal-bench-4-is-saturated/) and [OSWorld](https://epoch.ai/benchmarks/osworld-2-1/review) contain many broken tasks with impossible criteria and unreliable graders susceptible to false positives and negatives. Designing RL environments and tasks that can realistically challenge frontier models without artificial handicaps increasingly requires not only deep domain expertise but also engineering, operational, and research excellence.\n\n#### What makes us different\n\nSeeing the gaps in existing benchmarks and RL environments firsthand, we founded Epsilab to focus on quality and realism of RL data. We ground our training gyms in real-world data and expert judgement from top industry professionals, so that our environments capture precisely how enterprise work actually gets done.\n\nMaintaining this standard and level of fidelity is demanding and costly, and it is central to our thesis. Using our environments, frontier models and AI agents can hill climb on the right capabilities. \n\n<https://www.youtube.com/watch?v=LjYFbHWGPDQ>\n\n#### About us\n\n* Joanne is a 2x YC founder and previously built a healthcare AI startup Copper Health &amp; worked in investment banking and PE\n* Fengjiao is also a 2x YC founder, did RL research at MIT Media Lab, worked as an ML researcher at Vimeo, and built an AI comics generation startup Shortbread\n* Evan previously was CTO and cofounder of Shortbread, and did research on LLMs at UC Berkeley\n* Matthew worked as a quant infra engineer at Tower Research Capital, an HFT firm\n\n#### **Reach out**\n\n* Visit our website at <https://epsilab.com/>\n* We’re hiring! Email us at [founders@epsilab.com](mailto:founders@epsilab.com) or just say hi in general 👋","slug":"XGA-epsilab-rl-environments-for-enterprise-work","created_at":"2026-11-11T13:00:00.182Z","updated_at":"2026-11-17T22:19:26.555Z","total_vote_count":44,"url":"https://www.ycombinator.com/launches/XGA-epsilab-rl-environments-for-enterprise-work","share_image_url":"//bookface-static.ycombinator.com/assets/ycdc/yc-og-image-c440a0ad1dacfb86eeeb343717479cc54d256614449b4ef719977a0a451f8bc8.png","company":{"id":36974,"name":"Epsilab","slug":"epsilab-com","url":"https://www.epsilab.com","logo":"https://bookface-images.s3.amazonaws.com/small_logos/164ea33865ee89b7b5313a5260d5dec8510fb8ec.png","batch":"Fall 2026","industry":"B2B","tags":["Machine Learning","Reinforcement Learning","B2B","Data Labeling","AI"],"search_path":"https://bookface.ycombinator.com/company/36974"}}