Safeguards Enforcement Analyst, User Well-being

Anthropic; Washington (District of Columbia); Hybrid; USD 245,000 - 285,000; Full time

Benefits offered by this job

Equity donation matching

Flexible working hours

Generous vacation and parental leave

Office space

Job summary

Anthropic is seeking a Safeguards Analyst for the User Well-being team. You will design and deploy mental health guardrails, iterate on detection systems, manage review queues, and evaluate interventions.

This role involves translating clinical guidance and legal constraints into concrete policy enforcement actions, tackling harms like suicide and self-harm. The team will collaborate with engineers, data scientists, product, and legal to ensure safe interactions with Claude and related products.

Qualifications

• Experience in trust & safety, product policy, content moderation, or a related field with exposure to mental health, suicide, self-harm, or related well-being harm areas.
• Experience designing or running experiments to determine intervention efficacy.
• Ability to translate policy definitions into measurable rubrics, guidelines, or criteria for human reviewers or automated systems.
• Experience managing or coordinating content review operations, including QA and workflow management.
• Proficiency in SQL and/or other data analysis tools to measure intervention efficacy and surface policy gaps.
• Experience working with generative AI products and writing prompts for content review, classification, or evaluation.
• Ability to turn open questions and data into concise and insightful analyses.
• Experience identifying risks and communicating findings to cross-functional stakeholders.
• Understanding of scaling product policies in content moderation.

Responsibilities

• Support the design and execution of interventions, defining key metrics and curating evaluation datasets.
• Partner with Engineering and Data Science to build, tune, and validate detection models for automated intervention systems, including threshold setting and precision-recall tradeoffs.
• Monitor interventions and detection systems performance over time.
• Review flagged content to drive enforcement and policy improvements.
• Support in-product features that connect users to crisis resources, coordinating with Product, Legal, and external partners on referral pathways.
• Provide detailed feedback on policy gaps to the Safeguards Policy Design team based on real scenarios.
• Stay up to date with AI policy developments and research to inform decision-making and workflows.

Skills

Trust & safety

Policy enforcement

Data analysis

Cross-functional collaboration

Experiment design

Content review operations

Generative AI products

Communication

Ambiguity judgment

Education

Bachelor's degree

Job description

About Anthropic

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems.

About The Role

As a Safeguards Analyst on the User Well-being team, you will be focused on supporting the design and deployment of mental health guardrails – iterating on detection systems, managing review queues, evaluating new interventions, and monitoring existing ones. Interventions range from steering how Claude responds in the conversation itself to in‑product features that connect users to other resources. This work involves translating expert clinical guidance, internal data analyses, and legal or engineering constraints into concrete changes to how we detect, review, and respond.

This team covers a broad set of interconnected harms including suicide, self‑harm, disordered eating, AI sycophancy, and emotional dependence on AI. This position may expand into broader areas of user well‑being enforcement over time. Safety is core to our mission, and you'll help shape policy enforcement so that our users can safely interact with and build on top of our products in a harmless, helpful, and honest way.

• Important context for this role: In this position you may be exposed to and engage with explicit content spanning a range of topics, including those of a sexual, violent, or psychologically disturbing nature.

Key Responsibilities

• Support the design and execution of interventions, defining key metrics, and curating evaluation datasets
• Partner with Engineering and Data Science teams to build, tune, and validate detection models for automated intervention systems, including threshold‑setting and precision and recall tradeoffs
• Monitor how interventions and detection systems perform over time
• Review flagged content to drive enforcement and policy improvements
• Support the development of in‑product features that connect users to crisis resources, working with Product, Legal, and external partners on referral pathways and user‑facing content
• Support the Safeguards Policy Design team by providing detailed feedback on policy gaps based on real scenarios
• Keep up to date with emerging AI policy and external research on AI's relationship to mental health, and use these to inform our decision‑making and workflows

Minimum Qualifications

• Experience in trust & safety, product policy, content moderation, or a related field, with direct exposure to mental health, suicide and self‑harm, or related well‑being harm areas
• Experience designing or running experiments, evaluations, or measurement studies to determine whether an intervention worked
• Experience translating policy definitions into measurable form — the rubrics, review guidelines, or classification criteria, whether applied by human reviewers or automated systems
• Experience managing or coordinating content review operations, including quality assurance and workflow management
• Proficiency in SQL and/or other data analysis tools to measure intervention efficacy, monitor workflow health, and surface policy gaps
• Experience working with generative AI products, including writing effective prompts for content review, classification, or evaluation
• Experience turning open questions and data into concise and insightful analysis
• Experience identifying emerging risks and communicating findings to cross‑functional stakeholders
• Understanding of the challenges involved in implementing product policies at scale in the content moderation space
• Sound judgment in ambiguous, high‑consequence cases, and comfort making a call and escalating appropriately when the available signal is incomplete

Preferred Qualifications

• Subject matter expertise in mental health, whether developed in academia, clinical practice, crisis intervention, trust & safety, or other related settings
• Experience building or evaluating LLM‑based classification systems
• Experience using agentic tools (e.g. Claude Code) to scale analysis or automate recurring work
• Experience working within crisis support

Annual Salary

$245,000—$285,000 USD

Logistics

Minimum education: Bachelor’s degree or an equivalent combination of education, training, and/or experience

Required field of study: A field relevant to the role as demonstrated through coursework, training, or professional experience

Minimum years of experience: Years of experience required will correlate with the internal job level requirements for the position

Location‑based hybrid policy: Currently, we expect all staff to be in one of our offices at least 25% of the time. However, some roles may require more time in our offices.

Visa sponsorship: We do sponsor visas! However, we aren't able to successfully sponsor visas for every role and every candidate. But if we make you an offer, we will make every reasonable effort to get you a visa, and we retain an immigration lawyer to help with this.

Guidance on Candidates' AI Usage

Learn about our policy for using AI in our application process.

How We’re Different

We believe that the highest‑impact AI research will be big science. At Anthropic we work as a single cohesive team on just a few large‑scale research efforts. And we value impact — advancing our long‑term goals of steerable, trustworthy AI — rather than work on smaller and more specific puzzles. We view AI research as an empirical science, which has as much in common with physics and biology as with traditional efforts in computer science. We're an extremely collaborative group, and we host frequent research discussions to ensure that we are pursuing the highest‑impact work at any given time. As such, we greatly value communication skills.

The easiest way to understand our research directions is to read our recent research. This research continues many of the directions our team worked on prior to Anthropic, including: GPT‑3, Circuit‑Based Interpretability, Multimodal Neurons, Scaling Laws, AI & Compute, Concrete Problems in AI Safety, and Learning from Human Preferences.

Come work with us!

Anthropic is a public benefit corporation headquartered in San Francisco. We offer competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and a lovely office space in which to collaborate with colleagues.