Health AI, built to be trusted.
I build and validate AI for health, and I borrow heavily from how living systems already solve the same problems. Physician turned builder (Mayo MD, biomimicry MS); most of my time now goes to agents, clinical ML tooling, and evaluation: the unglamorous work that decides whether any of this is safe to act on.
- Atlas is an always-on Claude Code agent that lives on your Mac and answers from your phone over Telegram. No API key, nothing leaves the machine.
- Health Agent Rounds sends a health AI agent "on rounds," an evaluation suite that puts clinical agents through realistic cases before real patients do.
- MLRx is clinical ML prototyping with rigor by construction: leakage-safe pipelines, the right metrics, Model Facts labels.
- The Long Listen is an interactive essay on reading the body's quiet, ambient signals.
I also advise EpiFocus (seizure prediction from brain connectivity) and the Pitt HexAI Lab. Earlier: FDA-cleared neuroimaging at Cortechs.ai, and clinical validation and safety at Scienza Health.
Website · Writings · Talks · Projects · LinkedIn · X · Scholar
Building or validating health AI and want to compare notes? Grab a slot. Based in Arizona. 🌵