Architect secure AI systems, evaluate Frontier AI models, and advance multi-agent safety evaluation and governance frameworks.