Program Manager, Safeguards Policy, Enforcement, and Threat Intelligence

Written by the T&S Network from Anthropic's posting.

Program management for the Policy Design, Enforcement, and Threat Intelligence teams in Anthropic's Safeguards org, which sets and enforces the limits on how Claude can be used. The role keeps cross-team workstreams across harm areas on track, coordinates launch readiness, and runs the process for external publications such as threat reports.

What you'd do

• Own cross-functional program management across the three Safeguards teams and their harm-area workstreams.
• Set quarterly plans and priorities, track status, and log who decided what and why.
• Line up partner teams in research, product, engineering, legal, and comms so dependencies surface early.
• Drive policy, enforcement, and threat intelligence readiness for model launches and product releases.
• Manage the review and approval process for blog posts, threat reports, and other external publications.
• Identify risks, blockers, and resourcing gaps, and escalate them with a proposed path forward.
• Build lightweight processes and tools, and retire those that no longer earn their keep.

What it asks for

• Track record of delivering on time across teams you don't manage, in a fast-moving technical setting.
• Ability to turn ambiguous goals into plans, owners, and milestones and keep them current.
• Strong written communication for status updates, decision records, and executive summaries.
• Experience running approval workflows with many stakeholders, such as publications or launches.
• Comfort with sensitive topics and high standards for confidentiality.
• Bachelor's degree or equivalent combination of education and experience in a relevant field.
• Program management in T&S, product policy, enforcement ops, threat intel, or AI safety (preferred).

Worth knowing

• Exposure to explicit sexual, violent, or psychologically disturbing content.
• Late nights or weekend work may be required.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Anthropicnorth_east before you apply. If we got something wrong, report this listing.