Compute Foundations Engineer — Kubernetes & GPU Infra

OpenAI; San Francisco (CA); On-site; USD 255,000 - 490,000; Full time

Job summary

OpenAI is seeking a Software Engineer for Compute Foundations in San Francisco to help manage GPU compute infrastructure across sites and data centers. You will design, build, and operate Kubernetes-based controllers and services that coordinate hardware, firmware, and software across clusters, while ensuring reliability as the fleet grows and evolves.

The role blends software architecture with on-prem and cloud hardware knowledge, offering a path to impact OpenAI’s research and products through

Qualifications

• Experience designing, implementing, and owning production distributed systems or infrastructure services.
• Experience developing infrastructure systems using Kubernetes APIs and reconciliation to manage resources.
• Understanding how bare-metal nodes move from power-on to a configured, workload-ready state (PXE, DHCP/DNS, BMCs, firmware, Linux, drivers).

Responsibilities

• Design, build, and operate Kubernetes-based controllers and distributed services coordinating infrastructure across sites.
• Define APIs and resource models for lifecycle operations across hardware platforms and providers.
• Build provisioning and configuration services coordinating network boot, hardware management interfaces, and deployment of firmware, OS images, drivers, and host config.
• Develop lifecycle management for discovery, allocation, provisioning, upgrades, maintenance, recovery, and decommissioning with health validation integration.
• Design reliable reconciliation and recovery handling concurrent changes and partial failures with staged rollouts.
• Improve control-plane throughput and API latency while respecting site system and provider rate limits.
• Build integrations bringing new sites and GPU hardware generations into the platform with cross-team collaboration.

Job description

OpenAI is seeking a Software Engineer for Compute Foundations in San Francisco to help manage GPU compute infrastructure across sites and data centers. You will design, build, and operate Kubernetes-based controllers and services that coordinate hardware, firmware, and software across clusters, while ensuring reliability as the fleet grows and evolves.

The role blends software architecture with on-prem and cloud hardware knowledge, offering a path to impact OpenAI’s research and products through