AI Safety Evaluations Analyst

Anthropic; New York (NY); Hybrid; USD 230,000 - 270,000; Full time

Benefits offered by this job

Competitive compensation

Equity donation matching

Generous vacation

Parental leave

Flexible working hours

Office space for collaboration

Job summary

Anthropic is seeking a Safeguards Enforcement Analyst focused on Safety Evaluations to join our cross-functional Safeguards team in a hybrid role. You will run evaluations, interpret results, and drive mitigations across policy, engineering, and product teams to ensure model safety and policy compliance.

Responsibilities include developing new evals, maintaining rigorous documentation, and building workflows that scale this work as Anthropic expands.

Qualifications

• Experience in trust and safety, content operations, policy enforcement, or related operational role at a tech company.
• Ability to navigate ambiguity and coordinate across teams to drive work to completion.
• Experience building processes, workflows, or programs from scratch.

Responsibilities

• Run evaluations, monitor results, and surface regressions to stakeholders.
• Coordinate creation of new evals and ensure alignment with evolving policies and model capabilities.
• Manage evaluation outcomes and drive mitigations as needed.
• Design eval quality processes and frameworks to keep evaluations high-signal as models evolve.
• Create product-specific evaluations as Anthropic's offerings expand.
• Design tooling improvements for self-serve eval creation.

Skills

Trust and safety

Policy enforcement

Program management

Cross-functional collaboration

Data tools (SQL)

Education

Bachelor's degree

Tools

SQL

Dashboards

Job description

Anthropic is seeking a Safeguards Enforcement Analyst focused on Safety Evaluations to join our cross-functional Safeguards team in a hybrid role. You will run evaluations, interpret results, and drive mitigations across policy, engineering, and product teams to ensure model safety and policy compliance.

Responsibilities include developing new evals, maintaining rigorous documentation, and building workflows that scale this work as Anthropic expands.