Anthropic
Product Manager, Safeguards (Account Integrity & Abuse)
$385K–$460K
San Francisco, CA +2 · Manager · Full-time · Posted Sep 19, 2026
- Location
- San Francisco, CA
New York City, NY
San Francisco, California, United States - Experience
- Minimum years correlate with internal job level; preferred: 5+ years in product management.
- Education
- Bachelor’s degree or an equivalent combination of education, training, and/or experience; relevant field demonstrated through coursework, training, or professional experience.
- Visa sponsorship
- Visa sponsorship is available, but not guaranteed for every role or candidate; the company will make every reasonable effort to obtain a visa after an offer.
About the job
Owns the design, development, and deployment of safeguards systems and product experiences to detect and mitigate risks in Anthropic’s AI models and products.
What you'll do
- Plan safety-by-design and downstream defenses across AI models, products, and cloud platforms.
- Develop safety evaluations, interventions, and metrics to assess and mitigate deployment and user risks.
- Set product direction and priorities, making technical and business tradeoffs and collaborating across research, policy, enforcement, and engineering.
What they ask for
- Technical judgment and experience designing or building safety systems with cross-functional teams.
- Ability to design metrics for risk, system performance, and user impact, and use them to make tradeoffs.
- Experience planning, building, launching, and measuring products or systems in a zero-to-one environment.
- Strong written and verbal communication of complex technical concepts to non-technical audiences.
Nice to have
- 5+ years in product management, preferably involving data, detection and interventions, infrastructure and tools, or evaluations.
Pay
USD 385,000–USD 460,000 base salary
Tools
Claude.ai, 1P API
AI focus
Build safeguards for frontier AI models and products, including safety evaluations, risk detection, interventions, and metrics to measure and mitigate deployment and user risks.
Model-written summary · the original posting is authoritative