Field Notes
Perspectives from someone who builds, not just advises. Everything here is grounded in shipped products, real deployments, and clinical environments that don't forgive bad assumptions. Synced from LinkedIn. Updated as the work evolves.
Reflections from ESTRO 2026 on the behind-the-scenes work that makes clinical AI trustworthy and deployable.
A citable technical report describing the design, implementation, and evaluation of a local-first, privacy-preserving personal health AI architecture.
What I took away from ASTRO 2025: the community isn’t waiting for AI. It’s pushing for readiness, trust, and real impact.
Why planning is the right starting point for RadOnc AI, and what clinics should do now to be ready for what’s coming.
A fun hands-on build: a lightweight AI helper that revived my garden on a tiny budget, plus what it taught me about shipping fast.
Curated data helps, but real-world messiness matters. A practical take on how data choices shape model behavior and trust.