Question everything. Build better.
Methods, experiments, reproducibility and lessons from the frontier of intelligent systems.
CURRENT FOCUS
Safer autonomous systems through practical security testing.
Research themes include identity for agents, tool authorization, runtime observability, retrieval security and reproducible LLM evaluation.
AI Agent IdentityMCP SecuritySecurity Evaluation
01Agent identity at enterprise scaleResearch direction
02Runtime authorization for agent toolsResearch direction
03Local LLM phishing evaluationResearch direction