Ankit Gupta Joins YC as General Partner, Bringing Deep ML Expertise
Y Combinator Garry Tan
YC just made Ankit Gupta a General Partner. He built an AI drug-discovery startup that Ginkgo bought — now he mentors AI founders full-time.
Y Combinator has a new General Partner, and this one comes with a résumé that reads like a checklist for anyone trying to build a serious AI company. Ankit Gupta spent the last several batches as a visiting partner, quietly working with dozens of founders before YC decided to make it official. He's not a career VC parachuting in from a hedge fund. He's someone who actually shipped machine learning models for a living, which in a moment when every accelerator claims to be an AI accelerator, counts for something.
Gupta's own founder story started in YC's Winter 2018 batch, when he co-founded Reverie Labs, a biotech outfit that used machine learning to design small molecules for drug discovery. That's a much harder problem than slapping a chatbot on top of a product. Reverie signed partnerships with pharmaceutical companies and pushed some of its own drug candidates forward before Ginkgo Bioworks acquired the company in 2024. Six years from founding to exit is not a blitzscale story, but it's a real one, with real science behind it.
Before any of that, Gupta picked up a bachelor's and master's in computer science at Harvard and published research at ICML, one of the more competitive venues in machine learning, on deep learning and large-scale model training. So he's sat on both sides of the table: writing papers about how to train big models, then actually building a company that depended on those models working in production against biological reality, which does not care about your loss curves.
YC is also using this hire to reopen a Cambridge, Massachusetts presence, the city where the accelerator originally started before Silicon Valley swallowed it whole. Gupta will be based there, commuting to San Francisco rather than the other way around. That's a small but pointed signal that YC wants a foothold near MIT and Harvard's research pipeline, not just Bay Area hustle. For founders building at what YC calls the bleeding edge of AI, having a partner who has actually trained models and sold a company built on them is a different kind of mentor than one who has only read the pitch decks.
My take
Every accelerator wants an AI partner now, but most of them are marketing hires dressed up in technical language. Gupta actually has the scar tissue — a real ML research background plus a startup that lived or died on whether its models worked, which got bought rather than shut down. Betting on people who've built and sold something in a hard technical domain, instead of people who just talk fluently about transformers, is the correct move, and more accelerators should copy it instead of chasing whoever tweets loudest about AGI.
Read more about this at: Y Combinator
