Safeguards Enforcement Analyst, Safety Evaluations

About the role

Anthropic’s Safeguards team enforces policies, protects users, and prevents misuse of its AI platforms. As a Safeguards Enforcement Analyst focused on Safety Evaluations, you will ensure models meet safety and policy standards before and after launch, running evaluations, monitoring results, and driving mitigations.

Key responsibilities

• Run and monitor model evaluations, interpret results, and surface regressions or unexpected behavior to stakeholders.
• Partner with policy and domain experts to design, create, and update evaluations aligned with evolving policies and threats.
• Coordinate cross‑functional teams to manage evaluation outcomes, drive mitigations, and translate findings into model improvements.
• Develop processes and frameworks for product‑specific evaluations as Anthropic’s product surface expands.
• Design tooling improvements to support self‑serve evaluation creation for non‑technical users.
• Maintain rigorous documentation for evaluation creation, execution, and interpretation.

Required profile

• Experience in trust & safety, content operations, policy enforcement, or a similar operational role at a tech company.
• Comfort working in ambiguous, fast‑moving environments and building processes from scratch.
• Strong program‑management instincts, able to track timelines, dependencies, and deliverables across multiple stakeholders.
• Ability to manage several concurrent workstreams, prioritize effectively, and switch contexts quickly.
• Clear, concise communication skills, both written and cross‑functional.

Required skills

• Proficiency with data tools such as SQL.
• Experience using spreadsheets for workflow management.