Applied Compute: three OpenAI alumni, $80M, and enterprise custom models

ypatil125 · x · 2026-09-29

Part of a series on Lenny's top 100 startups, this profile covers Applied Compute: founders Yash Patil (23, Stanford dropout who worked on post-training and Codex at OpenAI), Rhythm and Linden, all OpenAI alumni. Their insight: OpenAI trains via evals/hill-climbing, and every company has proprietary tasks and data needing custom models. The company launched publicly in October 2025 with $80M in funding and early customers including Cognition, DoorDash and Mercor, aiming for Palantir-style enterprise AI.

Original post →

More from Companies & People

Companies & People channel →