Hi there... 👋

My home turf is a decade of computer science, malware analysis, and ML internals. The through line now is a single arena: securing the systems and the models we are all starting to build on. On the defensive side I build Aegis, a security kernel that hardens AI agents against prompt injection, tool shadowing, and memory poisoning. On the offensive side I run an abliteration lab, taking apart refusal directions and red-teaming model internals. Attack and defence on the same frontier, because you cannot secure what you have not learned to break.

The other half of the job is writing it down. I edit The Malware Files and publish long-form technical work on freeCodeCamp and elsewhere, a footprint that clears roughly 773,000 reads a year. Every piece ships with the methodology, the tooling, and enough detail for a reader to reproduce it. Reproducible or it did not happen.