The opportunity
About the Team The GPT Infrastructure team builds systems that turn advances in model inference and optimization into reliable production capabilities. We enable OpenAI workloads to be qualified and optimized across new accelerator platforms without requiring a one-off port and tuning effort for every hardware target.
What you'll do
Design, build, and operate APIs and control-plane services for long-running: workload qualification and optimization campaigns, including scheduling, retries, checkpointing, resource budgets, and observability.
Build secure partner-side execution and evaluation software that can compile,: run, verify, profile, and benchmark candidate artifacts on accelerator hardware.
Integrate model workloads, hardware profiles, compiler toolchains, runtimes,: serving engines, and distributed-execution backends into a repeatable platform.
Develop correctness and performance evaluation systems spanning output: fidelity, latency, throughput, memory footprint, accelerator utilization, communication efficiency, scaling behavior, and cost efficiency.
Automate the generation, evaluation, and improvement of kernels, runtime: configurations, parallelization strategies, and serving-stack changes.
Diagnose performance and correctness issues across model code, kernels,: compilers, runtimes, memory systems, networking, collective communication, and hardware.
What they're looking for
- Build artifact-management, provenance, regression-testing, and qualification: workflows for kernels, binaries, configurations, evaluation results, and deployment reports.
- Turn experimental research workflows into reliable product surfaces with: clear interfaces, actionable failure modes, and strong developer ergonomics.
- Collaborate with Research, Inference Engineering, Runtime and Compiler teams,: Infrastructure, Security, Product, and Strategic Partnerships to onboard and optimize new compute platforms.
- Drive technical architecture and execution across ambiguous initiatives: spanning OpenAI systems and partner environments.
