hey! my name is

Aadi Shah

about me:

  • I study computer science and philosophy at Stanford. Some of my favorite classes so far include CS224R (Deep RL), PHIL2 (Moral Philosophy), and CS231N (Deep Learning for Computer Vision).
  • I'm part of the Stanford AI Lab, where I work on robot learning with the IRIS and ILIAD groups advised by Chelsea Finn and Dorsa Sadigh. Recently, I've been working on memory for long-horizon manipulation tasks.
  • This summer I'm interning at Google on the Applied AI team, working on improving agent behavior for multi-turn conversations.
  • Previously, I led engineering at Mylon, an early-stage startup building AI-powered tools for research and knowledge discovery.

some projects:

  • For CS224R, my team built Ψ₀-Recap— a from-scratch port of Physical Intelligence's RECAP method onto the Ψ₀ humanoid VLA. RECAP lets the robot improve from its own rollouts by conditioning the policy on high-advantage actions.
  • I co-authored AutoSimTest, a paper on using LLM agents to automate drone simulation testing. The agents help write scenarios, run sims, and figure out what went wrong.
  • At CalHacks 12.0, we built krates, your AI devops engineer before you can afford to hire, automating Docker and Kubernetes for early-stage teams. krates won 1st place on the YC Track.

Last updated: July 2026

aadi2 [at] stanford [dot] edu