OpenAI posts Networking OS firmware role

OpenAI added a Networking Operating System Firmware Engineer listing to its careers site in San Francisco, according to a job page available on May 22. The posting places the role inside OpenAI’s Hardware organization, which the company says is building silicon and system-level solutions for advanced AI workloads. The listing says the team works on “the next generation of AI-native silicon” and partners with software and research groups on tightly integrated hardware systems. The new opening adds to a cluster of networking and systems roles that are already live in San Francisco. OpenAI’s careers site also lists openings for Software Engineer, Core Network Engineering; Software Engineer, Productivity - Networking; and Systems Engineer (Network / Storage / Systems), among others. Those postings describe teams working on data-center networking, WAN and edge connectivity, host and xPU networking, and infrastructure that supports large-scale training and inference. (openai.com) ### What exactly does this firmware role cover? The job page says the engineer would “build and own firmware and software for OpenAI’s networking stack.” The listed scope includes “switches, NICs, host networking, and management systems,” indicating work across both embedded and host-side components rather than a single device layer. The role also calls for work on “performance, observability, and reliability” in networking systems used for AI infrastructure. (openai.com) San Francisco is the named location on the posting, and the role is marked full-time on OpenAI’s site. The team description ties the job to OpenAI’s Hardware organization rather than a general software group, which places it alongside other hardware and firmware openings such as ASIC Firmware Engineer, Modeling and System Software Engineer, First-Party Hardware. ### Why is a networking operating system hire notable? (openai.com) OpenAI’s own job descriptions show a broad push into lower-level infrastructure roles tied to compute, networking and custom hardware. The Core Network Engineering posting says that team owns the “end-to-end networking stack” connecting OpenAI’s compute infrastructure across global WAN, data-center networking and high-performance host and xPU networking. The networking productivity role separately says its users build and operate networking systems that support training and inference “at frontier scale.” (openai.com) The hardware postings use similar language about co-designing hardware for AI workloads. The ASIC firmware and system software roles both sit in the same Hardware organization and describe work on AI-native silicon and system-level solutions, suggesting the networking firmware opening is part of a wider build-out rather than a standalone hire. ### What can be verified about the Codex usage claims? An OpenAI forum post published in December 2025 said Codex adoption among technical staff was above 92%, that Codex reviewed every internal pull request, and that engineers using it shipped about 70% more merged pull requests. (openai.com) That post did not use the exact figures circulating on X that said 95% of engineers use Codex and review time fell to two to three minutes, but it does support the broader claim that OpenAI has been publicly citing internal Codex adoption and pull-request gains. (openai.com) OpenAI’s Codex product and developer pages also describe code review as a core use case. The company says Codex can automatically review GitHub pull requests for regressions, missing tests and documentation issues, and its documentation includes setup instructions for automatic reviews and review triggers. ### Where does this fit in OpenAI’s current hiring picture? OpenAI’s careers homepage says the company is hiring across a wide range of disciplines and backgrounds. (forum.openai.com) On May 22, the San Francisco listings included hardware, firmware, networking and systems jobs alongside broader engineering roles. The Networking Operating System Firmware Engineer posting remains live on OpenAI’s careers site as of May 22, and applicants are directed to apply through that page. Nearby listings on the site show San Francisco as the location for related networking, hardware and systems roles that are still open. (developers.openai.com) (openai.com 1) (openai.com 2)