Technical Program Manager, Agent Safety & Security

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 We are hiring a Technical Program Manager to lead high priority safety and security programs which advance defense in depth for agents, which we have written about publicly . You will be the lead TPM coordinating work across teams to strengthen how we develop and deploy agents safely and securely, and to bring what we build internally to customers and partners so everyone has the tools to safely and securely use Claude. You will be responsible for leading meetings with executives and technical leads, maintaining complex project plans, identifying and addressing risks, developing metrics, and communicating to the company and our customers about this work. Responsibilities: • Maintain the roadmap - Maintain and extend the existing roadmap for all work contributing to this effort across the company and implement approaches to ensure accuracy and relevance amidst a fast operational tempo. • Turn program goals into measurable exit criteria at each phase of the work - To do this well you’ll need to be responsive to new developments in capabilities and the threat landscape, both of which are evolving fast. Maintain key metrics and dashboards. • Run the operating rhythm - Chair the leads meetings, working group meetings, design reviews, and executive check-ins. Record decisions and actions, disseminate sitreps quickly and regularly, and follow up until open issues close. • Lead executive and contributor communication - Drive conversations, write regular reports, contribute to design documents, and produce decisional documents as needed. Say plainly what is on track, what is not, and where help is needed. • Maintain the risk register - Track risks, exceptions, and risk acceptances in one place and bring hard calls to the right decision maker. • Drive company-wide communications - Partner with internal communications on progress updates and on rollouts that change how employees work with Claude every day. • Lead through other TPMs - Coordinate subject matter specific TPMs and engineering managers who do not report to you. Set shared standards, spot coverage gaps, and step in to augment capacity where needed. • Take coordination load off technical leaders - Find owners for cross-team problems, frame tradeoffs across Security, Safeguards, Research, Privacy, Legal, and Product, manage staffing tasks, and organize exercises that test our controls. • Maintain industry context - Track relevant incidents, capabilities shifts, and research across frontier labs, and understand the key risks they reveal. Bring those lessons into our threat model, roadmap, and risk register. You may be a good fit if you: • Have led company-wide technical programs, especially a company-wide security program, with multiple workstreams, executive sponsors, and other program managers rolling up to you. • Have a strong background in at least one of the program's domains: agent or infrastructure security, ML classifiers and monitoring, evaluations, inference infrastructure, protecting an organization’s most sensitive data, or shipping security controls to customers. You know the terminology and can ramp up on the rest quickly. • Have worked closely with ML research teams and understand how models are trained and evaluated. • Work independently. You set direction without waiting for instructions and know when to make decisions and when to escalate. • Can command a room and hold senior engineers, researchers,