Policy Design Manager, Conventional Weapons

$245k - $285k Remote-Friendly (Travel-Required) | San Francisco, CA | New York City, NY; Washington, DC

Posted 1d ago

Remote Work Policy

On-site

Categories

Applied AI Engineer

About the job

Anthropic's Safeguards organization is responsible for developing the policies, evaluations, and enforcement systems that define and limit how Claude can be used. In this role, you will own our conventional weapons work. This involves defining the boundary between acceptable and harmful requests in a domain where underlying components are dual-use, meaning they can serve civilian engineering and research while also contributing to weapons systems. The core of this role is building the threat models, evaluations, and detection systems to maintain this boundary. The role spans all weapon classes, including autonomous systems, and requires translating technical judgment into actionable principles for engineers and enforcement teams.

Responsibilities

  • Own and maintain Anthropic's conventional weapons policy, defining the boundary for model support.
  • Develop and update threat models and evaluations to measure model contributions to weapons development, including software and autonomy.
  • Collaborate with engineering to implement policy into model guardrails, detection systems, and enforcement tooling.
  • Act as the subject-matter expert for conventional weapons escalations and respond to emerging risks.
  • Build cross-functional understanding of the policy by communicating its reasoning to technical and non-technical audiences.
  • Engage external experts, government, and industry partners to strengthen policy and enforcement.

Requirements

  • Deep, applied expertise in weapons systems and ability to translate technical evidence into policy judgments.
  • Experience in a relevant setting (e.g., service research lab, defense research agency, weapons system company).
  • Ability to write clear, operationally precise policy and explain complex technical topics to non-specialists.
  • Understanding of legal frameworks governing weapons and their transfer.
  • Capability for rigorous technical analysis of weapons systems using open sources.
  • Comfort with ambiguity and complex problems without established playbooks.
  • Motivation to prevent misuse while not obstructing legitimate work.

Benefits

  • Annual compensation range: $245,000 - $285,000 USD
  • Visa sponsorship available

About Anthropic

Get new AI jobs in your inbox

A weekly digest of the newest AI engineering roles.

© 2026 AI Job Board. All rights reserved.