Writing
All
writing.
Notes from research and engineering practice, with a focus on trustworthy AI and real-world systems.
[ 7/4/2026 ]
When the Model Updates Itself: Regulatory Friction at the Boundary of Adaptive AI and Clinical Safety
The FDA cleared your model. Then it learned something new. Now what?
Read post
[ 6/25/2026 ]
Training on the Device, Not the Data: What On-Device Privacy-Preserving ML Actually Costs in Health AI
Everyone wants privacy-preserving ML. Fewer people want to talk about what you actually sacrifice to get it on a real device with a real patient's data.
Read post
[ 6/20/2026 ]
Autonomous Medical AI Agents Are Not Ready to Close the Loop
Nature just published a roadmap toward autonomous medical AI agents. Before we celebrate, let's be precise about what 'autonomous' means when a wrong decision can harm a patient.
Read post
[ 6/15/2026 ]
When the Clinical AI Forgets: Designing MAPE-K Loops for Model Drift in Pediatric Health Systems
Your pediatric brain health model was accurate at launch. Six months later, it is quietly wrong. Here is why static monitoring is not enough, and what a self-healing control loop actually looks like in practice.
Read postArchive