Product Manager, Safeguards Rare Harms
Core
Designing and deploying safety systems, evaluations, and interventions to protect users from risks associated with Anthropic's frontier AI models and generative AI products.
Role type
Product Manager, AI Safety & Safeguards
Builds
Safeguards systems, detection tools, evaluation frameworks, and product UX for AI safety across Claude.ai, 1P API, and external cloud providers.
Domain
Artificial Intelligence / AI Safety / Generative AI
Deliverable
production ML models | product features | dashboards & analysis
Required skills
Product management, safety system design, technical tradeoff decision-making, roadmap planning, metrics development, cross-functional collaboration, risk mitigation strategy, ambiguous environment navigation, zero-to-one product building, complex technical communication.
Preferred skills
Experience with AI/ML research teams, policy expertise, data detection and intervention infrastructure, user understanding of AI product usage.
Technologies
AI safety frameworks, evaluation tools, detection systems, cloud platforms, API infrastructure.
Responsibilities
Define upstream safety-by-design strategies and downstream defenses for frontier models; write safety evaluations and communicate safety findings externally; prioritize safety initiatives by defining problems and solution options; align with policy, research, engineering, and cross-functional stakeholders; plan for mitigation of deployment risks from powerful models; develop metrics to measure system performance, blindspots, and user impact.
Seniority
Mid-to-Senior, hands-on IC
