I build AI safety tooling that turns risk concepts into repeatable tests, guardrails, and operational monitoring. Three years empirical research on frontier AI models, focused on red-teaming, evaluation frameworks, and failure-first methodology. Seven years translating complex technical findings into actionable insights for government decision-makers.
Core Competencies
AI Safety & Evaluation
Red-teaming, adversarial testing, failure-first methodology
Frontier AI Models
Claude API, multi-agent systems, evaluation frameworks
Risk Assessment
Pre-mortem analysis, FMEA, failure mode identification
Policy Translation
Technical findings to actionable governance for decision-makers