Security Engineer | Adversarial AI Researcher | Marine Veteran
Recent: Built Aegis at Google I/O Hackathon 2026 (Judges' Round) — an autonomous purple-team multi-agent swarm powered by Gemini 3.5 Flash that runs concurrent adversarial mutations (OWASP LLM Top 10) against target agents and self-hardens prompt boundaries in a closed execution loop. Ships with real-time 3D telemetry.
Currently:
- Researching multi-agent adversarial systems and LLM red-teaming
- Building offensive security tooling and automated purple-team frameworks
- Sharpening on TryHackMe: gRRReat
Stack & Focus:
- Languages: Python, Bash, anything with a terminal
- Domains: LLM security (OWASP LLM Top 10), agentic systems, prompt injection, adversarial ML
- Mindset: Break it → understand it → harden it
Background:
- Former US Marine — two combat deployments to the Middle East
- Trusted with a TS clearance
- Disciplined operator translating mission-critical thinking into cyber defense
Off the clock:
- Magic cards are my favorite
- Been taking things apart since my first Compaq Presario in '98