Anthropic· Technical Program Management · San Francisco, CA | New York City, NY
Technical Program Manager, Research
Classified Tasks (23)
Automate 0%Augment 65%Human-Only 35%
Augment (15)
AI assists, human decides
Define and build programs that research teams need across the model development lifecycle
operational
Explore new research domains as priorities shift and determine each domain's specific needs
analytical
Identify highest-leverage opportunities and gaps for impact within research domains
analytical
Build processes that bring structure to unstructured research environments without slowing researchers
operational
Build frameworks and tooling that enable researchers to focus on research
technical
Establish playbooks and precedents where none exist to guide research initiatives
operational
Lead large-scale compute resource planning across research and production workstreams
operational
Allocate compute resources and prioritize workloads across research and production
operational
Optimize compute efficiency across research and production workstreams
technical
Standardize evaluation results to ensure consistent reporting for model launches
operational
Shape evaluation plans early in the model development lifecycle
analytical
Improve evaluation tooling to increase eval readiness for model launches
technical
Own execution and operational health of RL environments across major training runs
operational
Feed operational and RL environment insights back into roadmap and planning
analytical
Analyze technical tradeoffs deeply and present clear, actionable recommendations to research leadership
analytical
Human-Only (8)
Requires human judgment
Move fluidly across research areas (compute, evals, RL environments, emerging initiatives) and ramp to depth quickly
operational
Embed deeply within research domains to understand the technical landscape
technical
Build trust with researchers and technical leaders through sustained domain engagement
leadership
Drive end-to-end execution of complex, ambiguous research initiatives spanning multiple teams
leadership
Ensure honest and transparent reporting of evaluation results across research, product, and marketing
communication
Coordinate cross-team trade-offs during RL training runs
operational
Act as the connective tissue between research, engineering, and product teams to reduce chaos and accelerate execution
communication
Respond to incidents on short notice, including weekends, and participate in incident response activities
operational
Job description
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 Anthropic's research organization works across the full model development lifecycle, from pre-training and post-training to alignment, interpretability, and safety, each operating at the frontier of AI development. As a Technical Program Manager for Research, you'll define and build the programs that research teams need most. You'll move across research areas like compute, evals, RL environments, and emerging research initiatives, going deep enough in each to understand how researchers work and what they need. You'll identify where the biggest opportunities for impact lie, find the highest-leverage gaps, and build the programs, processes, and tooling that allow researchers to focus on research. This is a 0-to-1 role: you'll explore new domains as priorities shift, determine what each one needs, and create lasting impact where none existed before. Note: This role may require responding to incidents on short-notice, including on weekends. Responsibilities Embed deeply within a research domain to understand the technical landscape, build trust with researchers and technical leaders, and identify the highest-leverage problems to solve, knowing the surface area will shift over time as research priorities evolve Move fluidly across research areas like compute, evals, RL environments, and emerging research initiatives, picking up new domains quickly and getting to depth fast Drive end-to-end execution of complex, ambiguous research initiatives spanning multiple teams, often without established playbooks or precedent Establish processes and frameworks that bring structure to unstructured research environments without slowing researchers down Lead efforts like large-scale compute resource planning, including allocation, efficiency, and prioritization across research and production workstreams Drive eval readiness for model launches by standardizing results, shaping eval plans early, improving tooling, and ensuring honest, transparent reporting across research, product, and marketing Own execution and operational health of RL environments across major training runs, coordinating cross-team trade-offs and feeding insights back into roadmap planning Equip research leadership to make decisions quickly by going deep on technical tradeoffs and presenting clear, actionable recommendations Act as the connective tissue between research, engineering, and product teams to reduce chaos and accelerate execution You May Be a Good Fit If You Have a background in ML research or engineering with several years of experience building technical programs from scratch, ideally with hands-on exposure to training, evaluation, or large-scale distributed systems Are a fast learner who can ramp on unfamiliar technical domains quickly and contribute meaningfully to discussions with researchers Are resourceful, high-agency, and able to navigate ambiguity and shifting priorities to drive progress in fast-moving research environments Have a track record of operational ownership of complex technical systems, including monitoring, incident response, and performance optimization Can reason about technical tradeoffs at depth across model architecture, training infrastructure, evals, or compute efficiency, and translate them into clear decisions for leadership Have excellent stakeholder