At Alice.io, I design adversarial security evaluation environments and indirect prompt-injection scenarios for agentic AI systems, testing whether they can be manipulated into policy-violating behavior. I build and validate agentic evaluations using deterministic checks and LLM-judged rubrics.
Through my independent research, I built DragonGate, an autonomous security research system spanning 350+ bug-bounty programs, and earned $4K+ in autonomous bounty rewards. At ETXcyberops, I built AI blue-team automations and am developing a managed continuous external penetration-testing service; both remain at the prototype or development stage.

