Product Manager, Safeguards (Account Integrity & Abuse)

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

Product management for the Safeguards team at an AI company, owning design and deployment of safeguards systems and related product UX that protect users of frontier models across Claude.ai, the first-party API and external cloud platforms. The team builds detections, evals, interventions and tools to measure and mitigate deployment and user risks, including misuse of AI services.

• AI misuse and jailbreaks

What you'd do

• Define safety-by-design upstream and downstream defenses for frontier models and products across several surfaces.
• Take safeguards systems and their user-facing product experience from idea through shipping.
• Prioritize ruthlessly by defining problems, solution options and business and technical tradeoffs.
• Set MVP versus ideal-state requirements for detections, interventions and tools.
• Align policy, enforcement, research, engineering and other cross-functional stakeholders.
• Lead development of metrics covering risk areas, system performance and blind spots.

What it asks for

• Experience making technical tradeoff decisions across policy, AI/ML research and software engineering teams.
• Demonstrated experience designing and building metrics for risk, system performance and user impact.
• Experience planning, launching and measuring new products or systems in zero-to-one settings.
• Strong understanding of how the company's products are used and their safeguards concerns.
• 5+ years in product management with depth in data, detection, interventions, infrastructure, tools or evals (preferred)

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.