Safeguards Enforcement Analyst, Radiological & Nuclear Harms

Join Anthropic as a Safeguards Enforcement Analyst specializing in radiological and nuclear harm mitigation. This full-time, remote-friendly position is based in our San Francisco, New York City, or Washington, DC offices and offers the opportunity to protect AI systems from weaponization pathways while advancing the field of AI safety.

Role Overview

This position focuses on developing and maintaining safeguards against the misuse of generative AI for radiological and nuclear harm. You will serve as a technical and policy expert who reviews AI system interactions, determines the legitimacy of user intent, and drives continuous improvements to detection and mitigation systems. Working at the nexus of nuclear threat analysis and content enforcement, you will combine technical expertise with judgment to identify and prevent harmful use cases while supporting legitimate research and safety work. This role involves exposure to disturbing content spanning diverse subject matter, and includes shared on-call responsibilities with the policy and enforcement teams.

Core Duties

• Implement and refine enforcement protocols targeting radiological and nuclear misuse scenarios within our broader CBRNE risk framework.
• Build and oversee the end-to-end monitoring pipeline for nuclear and radiological threat detection, including triage, investigation, and response workflows.
• Identify emerging threat patterns and behavioral signals in platform activity that may necessitate policy refinements or immediate interventions.
• Architect scalable automated systems and human-review processes that balance precision with coverage across technically demanding domains.
• Execute detailed investigations into suspicious activities, compiling evidence and documentation to support enforcement determinations.
• Surface insights about detection gaps and recommend algorithmic or procedural enhancements.
• Collaborate with engineering and data science teams to refine detection algorithms and escalation systems.
• Work across Policy and Threat Intelligence teams to map exploitation vectors, develop risk frameworks, and advance safety-system development.
• Flag policy blind spots and manage time-critical issues involving radiological or nuclear policy breaches.

Requirements

• Bachelor's degree in physics, nuclear engineering, health physics, radiochemistry, or equivalent professional background in these domains.
• Hands-on experience in trust and safety, content moderation, or enforcement operations at scale, with demonstrated proficiency applying generative AI tools to optimize review and enforcement workflows.
• Proven ability to use AI tools for building data dashboards and tracking performance metrics in service of operational improvements.
• Capability to decompose ambiguous scenarios, synthesize evidence, and reach sound, defensible conclusions under time constraints.
• Initiative-driven mindset: you identify emerging issues, investigate independently, and implement solutions without waiting for direction.
• Capacity to engage with sensitive and potentially upsetting material sustainably over an extended period.
• Strong written communication skills and facility translating technical nuclear and radiological concepts across technical and non-technical stakeholder groups.

About Anthropic

At Anthropic, we are committed to advancing AI systems that are trustworthy, understandable, and aligned with human values. We believe AI development must prioritize safety and positive societal impact. Our diverse team—comprising scientists, technologists, policy leaders, and business experts—collaborates to ensure AI systems remain beneficial for the users and communities they serve.

Ready to Apply?

To apply, complete your application directly on this page, or you'll be redirected to the employer's application platform to finish submitting there.