- Build and scale red-teaming tools and products to improve the speed, coverage, and importance of findings.
- Operationalize systems for attacker simulation and output severity evaluation.
- Maintain and evolve vulnerability databases, statistical analysis tools, and reporting infrastructure.
- Develop agentic systems for automated red-teaming, jailbreak discovery and integration, and workflow automation.
- Build, test, and refine sandboxing architecture for evaluating advanced AI systems in high-stakes risk domains.
- Extend infrastructure and tooling for CBRN and cyber evaluations and loss-of-control testing.
- Design architecture for external platform contributions and integrations with the AI evaluation ecosystem, including Inspect.
- Build and manage an engineering team, mentor team members, and develop hiring pipelines.
- Translate division technical priorities into scalable engineering plans, teams, and systems.
PythonCybersecurityKubernetes+2 more