Technical Program Manager, Compute Infrastructure

ABOUT THE TEAM The Compute Infrastructure team runs the GPU fleet and large-scale compute clusters that serve the models backing ChatGPT and the API, while also supporting training workloads for our next generation of models. We provide a unified platform for OpenAI teams to run production Applied AI and Research training workloads. Our work spans cloud providers, bare metal infrastructure, data centers, and the foundational services that make compute usable. We partner across Engineering, Research, Product, Capacity Planning, Finance, and Data Center Infrastructure to bring new capacity online and support OpenAI’s growing compute needs. We seek to learn from deployment and distribute the benefits of AI, while ensuring that this powerful tool is used responsibly and safely. Safety is more important to us than unfettered growth. ABOUT THE ROLE As a Technical Program Manager for Compute Infrastructure, you will own the supply signal for all GPUs entering our fleet: a reliable view of what capacity is arriving, where and when it will arrive, and when it is expected to become usable. You will own the roadmap, requirements, delivery, and adoption of the tools, data, and platform behind that signal, partnering with Engineering to build systems that support planning and execution across OpenAI. You will also lead programs that turn incoming supply into usable compute. Working across cloud providers, Data Center Infrastructure, and compute engineering teams, you will connect physical site readiness with the CPU, storage, network, GPU, and foundational services required to bring new sites and clusters online. You will be part of an engineer-first TPM team, managing a broad portfolio of infrastructure programs and making dependencies, risks, and decisions clear from working teams through leadership. Your work will help teams plan with confidence, respond to supply changes, and deliver capacity on time. IN THIS ROLE, YOU WILL: - Own the GPU supply signal across the fleet. Maintain a trusted view of incoming capacity across providers, sites, and clusters, including quantities, delivery forecasts, acceptance and readiness milestones, and expected dates for production use. Make changes and their downstream effects visible. - Build and scale the tools, data, and platform behind that signal. Own the roadmap, requirements, delivery, and adoption, partnering with Engineering on data models, integrations, quality checks, and planning workflows. Serve Product TPMs, Capacity TPMs, and Compute TPMs, alongside Finance, Accounting, leadership, Engineering, and Research. - Connect supply with capacity planning. Partner with Capacity Planning on supply and demand management, allocation planning, and scenario analysis. Ensure plans reflect current supply forecasts, infrastructure constraints, and the effects of early or late deliveries. - Coordinate physical readiness for new capacity. Work closely with Data Center Infrastructure and cloud service providers to align ready-for-service dates, site construction, power and cooling readiness, supply-chain convergence, equipment delivery, and commissioning. Track dependencies and drive resolution of gaps that threaten deployment. - Lead logical readiness across infrastructure resources. Coordinate CPU, storage, network, GPU, Kubernetes, and dependent compute foundation services so new clusters can support their intended workloads. Make clear which missing services or resources could delay GPU use. - Drive deployment across cloud and bare metal environments. Lead integrated plans across AWS, GCP, Azure, OCI, and other infrastructure partners, accounting for provider-specific dependencies while establishing consistent readiness criteria and handoffs. - Establish Delivery Scorecards for cloud providers and internal infrastructure teams. Track delivery commitments, forecast changes, actual delivery, readiness milestones, and time to usable capacity. Use these scorecards to identify recurring gaps, clarify recovery actions, and improve delivery predictability. - Improve the reliability and repeatability of cluster bringup. Partner with Engineering to scale automation, runbooks, validation, and deployment processes. Reduce manual coordination and the time between capacity delivery and production use. - Manage the critical paths across concurrent programs. Establish clear milestones, owners, dependencies, and escalation paths. Resolve cross-team blockers and drive decisions that keep site and cluster deployments moving. - Communicate progress, risks, and tradeoffs clearly. Give technical and business stakeholders the information they need to understand capacity availability, assess changes to the plan, and make timely decisions. YOU MIGHT THRIVE IN THIS ROLE IF YOU: - Have 5+ years of experience leading major technical programs, including compute infrastructure, hyperscaler deployments, capital projects, or similarly complex infrastructure work. - Possess a degree in a hard science or engineering discipline, or have a demonstrated track record of equivalent engineering expertise. - Are familiar with Kubernetes deployments and have worked with at least one major cloud provider—such as AWS, GCP, Azure, or OCI—to scale compute infrastructure. Experience coordinating deployments across multiple providers is valuable. - Are familiar with bare metal GPU capacity and the provisioning, inventory, and planning systems used to bring it into service. - Have worked with Capacity Planning teams on supply and demand management, allocations, constraints, and supporting tools. You understand how changes in delivery or readiness affect downstream capacity plans. - Have experience defining requirements for operational data and internal platforms. You can work with engineers to establish useful data models, improve data quality, and build tools that teams adopt and rely on. - Understand how site readiness, power and cooling, equipment delivery, commissioning, and ready-for-servi