Anthropic
Threat Intel Manager, Model Exploitation & Fraud
$375K–$455K
San Francisco, CA +1 · Manager · Full-time · Posted Jul 8, 2026
- Location
- San Francisco, CA
San Francisco, California, United States - Experience
- Years required correlate with internal job level requirements; no specific number stated.
- Education
- Bachelor’s degree or an equivalent combination of education, training, and/or experience; field relevant to the role.
- Visa sponsorship
- Visa sponsorship is available, but not for every role or candidate; the company says it will make reasonable efforts to obtain a visa if it makes an offer.
About the job
Leads Anthropic’s Model Exploitation & Fraud team to detect, investigate, and disrupt large-scale abuse of its AI systems and translate findings into enforcement and product improvements.
What you'll do
- Set mission strategy, lead and develop technical investigators, and define casework quality and team metrics.
- Direct complex investigations and improve high-volume detection and triage, including abuse signals, clustering, and agentic workflows.
- Lead government and industry intelligence sharing and coordinate with policy, enforcement, and engineering on mitigations.
What they ask for
- Experience leading and managing investigative, fraud, platform integrity, or threat intelligence teams.
- Strong scaled-abuse domain fluency and experience overseeing investigations across surface, deep, and dark web environments.
- Proficiency in SQL and Python; working familiarity with large language models and model exploitation.
- Experience building processes, detection systems, or programs from scratch.
Nice to have
- Experience at a major technology platform in trust and safety, fraud, or abuse investigations; financial crime or fraud analytics; or direct work with U.S. government stakeholders.
- Mandarin Chinese and/or Russian fluency, or active Top Secret security clearance.
Pay
$375,000 — $455,000 USD (base salary)
Tools
SQL, Python
AI focus
Investigates model distillation, extraction, and exploitation; builds detection and agentic investigation workflows for large-scale abuse.
Model-written summary · the original posting is authoritative