The Problem
Unit tests. Static analysis. Code reviews. System testing. Then hundreds or thousands of pages documenting that you did all of it. Bugs still get through.
Early Results
Found real bugs in production code that passed human review.
What MedSafe-AI Does
AI analyzes your source code and finds issues traditional methods miss.
Requirements Verification
Given your requirements and code, MedSafe-AI verifies each requirement is correctly implemented. You get a per-requirement verdict and a traceability matrix.
Defect Detection
MedSafe-AI analyzes code against AAMI SW91 defect categories—buffer overflows, null dereferences, race conditions, unchecked returns, and more. You get findings with exact line numbers.
Your Code Stays Protected
Analysis runs locally. Your code never leaves my possession, never goes to the cloud. NDA available.
About
Mike Luloh — 25 years in medical device software, including roles at Alcon and as a voting member on the AAMI standards committee that shaped IEC 62304.
Why I'm doing this: AI will eventually verify software more thoroughly than humans can. (Naturally, embedded devices still need hardware integration testing.) When that happens, the current regime—hundreds of pages of documentation proving humans followed the process—becomes unnecessary. I'm gathering evidence now to prove it works.
This isn't an outside disruption. It's an insider who sees where this is heading—and knows the current system won't survive the transition.