Technical Program Manager, Global Infrastructure Planning

Job description

About the Team The Compute Infrastructure team runs the GPU fleet and large compute clusters that support OpenAI’s research, ChatGPT, and the API. Our work spans GPUs, CPUs, storage, networking, and cloud infrastructure, bringing these systems together so teams can train and serve models at scale. 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 You will own CPU capacity planning and management across Research and Compute. Working with infrastructure owners, you will lead integrated planning for CPU, storage, and network capacity across OpenAI’s global fleet. You will connect workload demand and GPU deployment plans with the foundational resources required to support pretraining, research, and application infrastructure at scale. You will work closely with workload owners, infrastructure engineering, network deployment teams, Finance, and cloud providers to turn forecasts into supply and allocation plans. You will understand workload architecture, guide decisions when capacity is constrained, and drive opportunities to improve utilization and reduce waste. Your work will span planning over multiple years and ongoing capacity management across regions and providers, including AWS, GCP, Azure, and OCI. This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees. In this role, you will: Lead CPU demand planning across Research, Compute, and application infrastructure, including foundational services, data acquisition, container platforms, robotics, and services supporting GPU clusters. Maintain forecasts that capture workload growth, regional and hardware requirements, timing, and acceptable alternatives. Translate workload growth and GPU fleet plans into integrated CPU, storage, and network requirements. Partner with engineering to determine the capacity, throughput, bandwidth, connectivity, and placement needed to support pretraining and applications at scale. Reconcile current and incoming supply with demand, model constrained scenarios, and drive allocation and prioritization decisions. Lead CPU allocation planning and fulfillment, including quota transfers, rebalancing, and reclaims. Partner with network engin…