Threat Intel Manager, Model Exploitation & Fraud

Anthropic•San Francisco, CA

7d•Hybrid

About The Position

Anthropic is seeking a Threat Intel Manager to establish and lead their Model Exploitation & Fraud team within Threat Intelligence. This role involves detecting, investigating, and disrupting large-scale exploitation of Anthropic's AI systems, including model distillation, unauthorized access, account farming, reseller abuse, and fraud/scam operations. The manager will define the strategic direction, recruit and lead a team of technical investigators, and develop scalable systems, processes, and partnerships. The team comprises senior investigators with deep technical expertise, and the manager's role is to direct, resource, and amplify their work. This position requires regular engagement with U.S. government entities and a thorough understanding of external black market ecosystems and their interaction with Anthropic's systems. Please note that this role may involve exposure to explicit content and may require responding to escalations during weekends and holidays.

Requirements

Led and managed investigative, fraud, platform integrity, or threat intelligence teams, ideally those built around senior, deeply specialized individual contributors.
Strong domain fluency in scaled abuse (fraud patterns, account abuse, unauthorized access, or platform exploitation economics) sufficient to set priorities, pressure-test findings, and earn the confidence of expert investigators.
Proficient in SQL and Python to review data-heavy casework, pressure-test conclusions, and provide surge capacity.
Experience overseeing investigations tracking threat actors across surface, deep, and dark web environments, including reseller and access-broker communities.
Working familiarity with large language models and a strong grasp of how models can be distilled, extracted, or exploited at scale.
Experience building processes, detection systems, or programs from scratch.
Ability to communicate crisply with executives, engineers, and external partners.

Nice To Haves

Experience at a major technology platform on trust and safety, fraud, or abuse investigations at scale.
Background in financial crime investigation or fraud analytics.
Experience working directly with U.S. government stakeholders on threat reporting.
A track record of partnering with, growing, and retaining senior technical specialists, including defining clear scope between management and senior IC tracks.
Fluency in Mandarin Chinese and/or Russian with nuanced regional and geopolitical context.
Active Top Secret security clearance.

Responsibilities

Own strategy, priorities, and outcomes for the Model Exploitation & Fraud mission area; define what we detect, investigate, action, and share.
Hire, manage, and develop a team of technical threat investigators; set the quality bar for casework and intelligence reporting.
Define clear lanes between the manager role and senior individual contributors: strategy, people leadership, and program ownership reside with the manager, while ownership of deep technical investigations and tradecraft remains with senior experts.
Independently lead complex investigations, directing, prioritizing, and resourcing investigations into model distillation, unauthorized AI R&D usage, unauthorized access, coordinated account abuse, and fraud/scam networks, while partnering with senior investigators and clearing blockers.
Drive the redesign of triage for a high-volume detection pipeline, partnering with investigators and engineering to build abuse signals, clustering, and agentic investigation workflows.
Expand the team's coverage into fraud and scams, building detection and investigation playbooks from scratch.
Own the external engagement program for the area, including intelligence sharing with U.S. government partners and industry peers, ensuring investigators driving the work are visible.
Anticipate how resellers, proxies, and third-party platforms impact the abuse surface and shape coverage accordingly.
Work with policy, enforcement, and engineering to convert findings into bans, product mitigations, and safety-by-design improvements.
Define and report the team's metrics; brief Safeguards and company leadership on the threat landscape.