Technical Program Manager, Core Network & WAN Infrastructure
OpenAI · San Francisco, CA · Technical Program Management
About this role
OpenAI is hiring a mid-level Technical Program Manager in the software engineering function based in San Francisco, CA. The posting calls out experience with Networking, Testing, Machine Learning, OpenAI. Compensation is listed at $207,000–$335,000 per year.
- Role
- Technical Program Manager
- Function
- software engineering
- Level
- mid
- Track
- Individual contributor
- Employment
- Full-time
- Location
- San Francisco, CA
- Department
- Technical Program Management
- Posted
- May 20, 2026
More roles at OpenAI
Job description
from OpenAI careersAbout the Team
The compute infrastructure team runs the GPU fleet and supercomputers that serve the models backing ChatGPT and API while also supporting training workloads for our next generation models. We manage one of the largest cutting edge GPU fleets in the world, exposing it as a singular platform for other OpenAI teams to seamlessly run production Applied AI and Research training workloads.
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
We're hiring a Technical Program Manager to own delivery of OpenAI's WAN infrastructure across PoPs, long-haul fiber routes, cloud interconnects, colocation environments, and provider handoffs.
This is a hands-on infrastructure execution role. You will drive PoP readiness end-to-end: equipment, cabling, optics, cross-connects, port maps, vendor dependencies, cloud connectivity, test plans, operational runbooks, and escalation paths. You should be comfortable working with network engineers, datacenter operators, cloud providers, fiber vendors, finance, procurement, and business teams, while maintaining crisp ownership of what is live, blocked, at risk, and needed next.
The right person has enough technical depth to reason through physical and logical network readiness, enough program discipline to keep complex builds moving, and enough ownership to improve how we scale instead of only tracking status.