Anthropic
Technical Program Manager, Agent Safety & Security
$365K–$435K
San Francisco, CA +2 · Manager · Full-time · Posted Oct 21, 2026
- Location
- San Francisco, CA
New York City, NY
San Francisco, California, United States - Experience
- 10+ years in one of the specified areas; the posting also says required years correlate with the internal job level.
- Education
- Bachelor’s degree or an equivalent combination of education, training, and/or experience; field relevant to the role, demonstrated through coursework, training, or professional experience.
- Visa sponsorship
- Visa sponsorship is available, but not for every role or candidate; if an offer is made, Anthropic will make every reasonable effort to obtain a visa and provides an immigration lawyer.
About the job
Leads company-wide safety and security programs to help develop and deploy AI agents safely and bring related tools and controls to customers and partners.
What you'll do
- Maintain the cross-company roadmap, measurable program criteria, metrics, dashboards, and risk register.
- Coordinate operating rhythms and decisions across technical teams, executives, and other TPMs; communicate progress and escalate issues.
- Coordinate cross-functional work, staffing, and control-testing exercises, incorporating industry incidents and research into program planning.
What they ask for
- Have led company-wide technical programs with multiple workstreams, executive sponsors, and other program managers.
- Have a strong background in at least one relevant domain, such as agent or infrastructure security, ML classifiers and monitoring, evaluations, inference infrastructure, sensitive-data protection, or customer security controls.
- Have worked closely with ML research teams and understand how models are trained and evaluated.
- Write clearly for varied audiences and use AI tools like Claude to automate tracking, reporting, and follow-up.
Pay
$365,000 — $435,000 USD (annual salary)
Tools
Claude
AI focus
Coordinate agent safety and security programs that strengthen safe development and deployment, including defense in depth and customer-facing tools and controls.
Model-written summary · the original posting is authoritative