About this role
Engineers in this group build distributed systems that provision, configure, and manage GPU compute infrastructure across data centers and providers to support model training and inference. Daily work involves designing lifecycle APIs, improving controller performance under high concurrency, and tracing provisioning failures from APIs through network boot or host configuration. Team members create Kubernetes-based control planes and services that turn heterogeneous fleets of machines into dependable compute resources.
OpenAI maintains Compute Foundations to connect global infrastructure management with the realities of bare-metal systems, ensuring consistent interfaces across varying hardware and provider behavior. This role exists to help these systems remain reliable as the fleet expands across sites and hardware generations. Professionals who combine software architecture with a practical understanding of how data centers and bare-metal machines operate will find success here.
What you will do
- Design and operate Kubernetes-based controllers and distributed services that coordinate infrastructure across multiple sites.
- Define APIs and resource models that let clients track lifecycle operations through consistent interfaces.
- Build provisioning services that handle network boot, hardware management interfaces, and software deployments.
- Develop lifecycle management for discovery, allocation, upgrades, and recovery integrated with health validation systems.
- Design reliable reconciliation and recovery mechanisms to limit disruption during partial failures.
- Improve control-plane throughput, API latency, and provisioning speed while respecting provider limits.
- Integrate new sites and GPU hardware generations into the platform by partnering with adjacent infrastructure teams.
What you need
- Demonstrate strong software engineering fundamentals with experience owning production distributed systems or infrastructure services.
- Build infrastructure systems that utilize Kubernetes APIs and reconciliation to manage resources.
- Understand bare-metal node startup processes with depth in areas such as PXE, DHCP/DNS, BMCs, firmware, or Linux drivers.
- Design reliable APIs and asynchronous workflows while reasoning about concurrency and failures across service boundaries.
- Diagnose performance problems across service, operating-system, and machine boundaries to deliver software improvements.
Nice to have
- Have built infrastructure control planes that coordinate operations across multiple sites or regions.
- Have worked with GPU or HPC infrastructure, including topology and shared dependencies.
- Have integrated multiple hardware platforms or infrastructure providers into a common service or resource model.
Benefits
- Offers Equity
Skills and tech stack
Questions candidates ask
What kind of systems does the Compute Foundations team build?
Compute Foundations builds the software that manages OpenAI’s GPU compute infrastructure across sites, data centers, and infrastructure providers. The team creates Kubernetes-based control planes, controllers, services, and APIs that coordinate the lifecycle of machines and clusters.
What will I actually work on in this role?
You will build distributed systems that provision, configure, and manage compute throughout its lifecycle. Your work connects global services and Kubernetes controllers with the systems that bring machines online, update them safely, and recover them when something goes wrong.
Do I need to be an expert in every part of the stack?
The team values depth in relevant systems and the ability to connect layers. You do not need to arrive as an expert in every component of the stack.
What is the compensation for this position?
The salary range is listed as $255K – $490K with offers of equity included. The position is based in San Francisco.