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
Pacing is Anthropic's work to make sure that, as AI capabilities advance, enough time and effort goes into making them safe. It covers the safety commitments we make and carry out ourselves, the public policy we develop and support, and the research and communication that help people understand how quickly AI is progressing and what that could mean.
We're hiring a Technical Program Manager to run Pacing. It spans many projects led by people across Safety, Research, Policy, Legal, Communications, and Engineering, most of whom have other responsibilities too. You'll make sure each project has an owner, a clear outcome, and a date, and that leaders have high visibility and can make tradeoff decisions. You'll work closely with the team responsible for our Responsible Scaling Policy and support every project lead in delivering on their commitments.
The questions and strategies relevant to this program will change rapidly as the frontier evolves. You'll need enough technical depth to drive discussions about model training, evaluations, and compute, and the judgment to identify which decisions belong with cross-functional partners and which you can make yourself. We're looking for a generalist who can ship quickly and iteratively and drive accountability towards high-level, ambiguous objectives.
Key responsibilities
Own the Pacing program end to end, keeping one accurate view of projects, owners, milestones, dependencies, and gaps that drives action from leadership and project leads
Turn strategy from executive sponsors into scoped projects with measurable outcomes and timelines for execution, and drive concrete proposals to decisions
Run the operating rhythm, including working sessions with project leads, regular written updates, and effective executive reviews, and follow decisions and action items through to completion
Work with the Responsible Scaling Policy team and researchers to define the evidence and sign-off a model needs at key points in training and deployment, and make those reviews a dependable part of how we work
Track the public and internal commitments Anthropic makes on pacing, and sequence the work, publications, and events behind them with Policy and Communications
Partner with Legal throughout the lifespan of each project, and operate with high integrity on confidential and sensitive work
Define how progress is measured, report against it candidly, and recommend which projects and practices to continue, change, or stop
Use Claude to build automations to allow teams to have broader and deeper impact
Minimum qualifications
Experience leading multiple complex technical programs that span organizations and involve engineering or research teams, from ambiguous goals through delivery
Strong program management skills, including program planning, milestone and dependency tracking, and risk management across many concurrent projects
Technical fluency sufficient to identify risks and challenge tradeoffs in how frontier AI models are trained, evaluated, and deployed
Working knowledge of AI safety and the main debates about how frontier AI should be developed and governed
A track record of delivering through senior leaders and teams who don't report to you, including when it meant changing how they work
Experience preparing decisions for executives, including framing options, making a recommendation, and recording the outcome
Sound judgment and discretion with confidential and sensitive information
Preferred qualifications
Experience at a frontier AI developer, a government AI institute, an independent evaluator, or an AI safety or governance research organization
Hands-on…