The opportunity
Frontier Systems Foundations, part of Compute Foundations at OpenAI, builds the systems software foundation that turns new compute infrastructure into reliable, usable capacity for frontier model training.
What you'll do
Build and maintain the Linux host software stack for large GPU clusters,: including Ubuntu and OS images, kernel configuration and modules, drivers, packages, disks and storage configuration, and machine configuration.
Design reproducible OS-image builds, package and repository workflows, and: system configuration for heterogeneous bare-metal and cloud compute fleets.
Integrate, test, and qualify kernels, modules, drivers, packages, and: firmware across new and existing hardware platforms; build safe canary, rollback, and recovery paths for system changes.
Bring up new hardware platforms and compute SKUs, working with hardware: engineers and vendors to resolve firmware, driver, operating-system, and compatibility issues.
Debug complex system failures across boot and provisioning, firmware, disks,: kernels and drivers, and workload interactions; turn recurring failure modes into durable fixes, tests, and automation.
Improve provisioning, repave, maintenance, and recovery correctness by: eliminating manual host-by-host intervention and making system behavior predictable at fleet scale.
What they're looking for
- Build focused systems tooling and diagnostics that make the host software: stack easier to validate, troubleshoot, and operate across the installed fleet.
- Have significant experience building, integrating, or operating production: Linux systems, particularly Ubuntu, Debian, or another major Linux distribution.
- Have deep experience in one or more of: Linux kernels, modules, or device drivers; Linux distribution, package, repository, or OS-image engineering; boot, provisioning, disks, or machine configuration; firmware and driver integration, qualification, or rollout.
- Can write, debug, and maintain production-quality systems software and: automation using appropriate systems languages, scripting, and Linux tooling.
