CEOInterviews.AI
Start App
Try: AI strategy tariffs earnings guidance layoffs interest rates growth

Yash Patil

Co-Founder and CEO, Applied Compute

Search every verified Yash Patil interview, podcast appearance, and on-the-record quote — each transcript cross-checked by AI and human review to confirm speaker identity. Yash Patil, co-founder and CEO of Applied Compute, discussed the company's focus on post-training and inference for open models in two recent podcast appearances. On the Unsupervised Learning podcast, Patil, a former OpenAI researcher who worked on Codex, argued that post-training is an overlooked part of the AI stack and that companies should prioritize building and safeguarding evals, describing reinforcement learning as a "hill climbing machine" where defining the objective is the hardest part. He also stated that inference, rather than training, accounts for the majority of model-related spending and that Applied Compute targets high-value use cases or large inference workloads. On the First Pass podcast, Patil discussed the launch of AC2, the Applied Compute Agent Cloud, and said that open source models combined with post-training can now deliver better models for specific use cases at lower cost, claiming the historical trade-off between quality and cost has changed. He noted that he and his co-founders left OpenAI to help companies "own their intelligence" and described Applied Compute's model factory approach, which includes tools for data analysis, ablations, and real-time reinforcement learning using production traces. Patil also advised that companies should first build evals, stating that most organizations are not "data ready" because measuring what they care about is difficult.

Recent appearances

  • Applied Compute CEO on the Limits of RL, the New AI Hyperscaler & Why Post-Training Wins Inference

    October 06, 2026 · 📺 Unsupervised Learning: With Jacob Effron

    Yash Patil, a former OpenAI researcher who worked on Codex and now runs Applied Compute, argues that post-training is the most overlooked part of the AI stack and the key to owning your intelligence. He explains when companies should optimize their harness and context first, and when it's worth updating the weights: the largest inference workloads, plus tasks that change over time or depend on judgment. That's the heart of why he believes post-training wins inference. The biggest workloads are both the most valuable to train and the most valuable to serve, and how a model is trained shapes ho…

    Watch on YouTube ↗

  • Are open models still a trade-off? | First Pass Ep. 10 with Yash Patil (Applied Compute)

    September 29, 2026 · 📺 Altimeter Capital

    Open models used to mean giving up quality to save money. Yash Patil says that trade-off is gone. Yash Patil, co-founder and CEO of Applied Compute, joins Jamin Ball and Palak Goel to break down AC2, the Applied Compute Agent Cloud, and why open source models combined with post-training can now deliver better models for a specific use case at lower cost. Yash also covers why he and his co-founders left OpenAI to help companies own their intelligence, why he believes the best customers already have well-defined evals, what an RL factory looks like when new open weights ship every week, why th…

    Watch on YouTube ↗

  • An Ex OpenAI Researcher on the Real Risk of Frontier Models

    June 23, 2026 · 📺 The Generalist

    Yash Patil is the 23-year-old founder and CEO of Applied Compute, a $1.3 billion company helping businesses train custom AI models on their own data: smaller, cheaper, and purpose-built for the work they actually do. Before founding the company, Yash dropped out of Stanford and spent two years at OpenAI working on post-training infrastructure and Codex. He left with one core conviction: every company that runs its critical workflows on someone else’s model is building on shifting sand. Applied Compute is his answer to that problem, already serving customers including DoorDash, Cognition, and M…

    Watch on YouTube ↗

  • Specific Intelligence at Scale: How Applied Compute Trains Custom Agent Workforces on Modal

    May 20, 2026 · 📺 Modal

    In this talk, we go behind the scenes with Yash Patil to look at the future of custom AI agents. Applied Compute trains custom AI agents for enterprises like DoorDash, Cognition, and Mercor. The founding team came out of OpenAI's Codex and o1 work, and started the company around a specific thesis: as frontier models commoditize, the competitive layer moves to post-training. Companies that own their reward functions, their evals, and their continual learning loops will pull ahead of companies that don't. They call this Specific Intelligence, and the full stack runs on Modal. Timestamps 0:00 —…

    Watch on YouTube ↗

The summary and quote tags on this profile are produced with AI assistance from verified, first-person interview transcripts, then checked by our team to confirm the speaker's identity and the accuracy of every quote. See how we verify →