Software Engineer, Inference - Performance Optimization

Software Engineer, Inference - Performance Optimization at OpenAI. This role focuses on analyzing and optimizing inference stack performance across application, model, and fleet layers to achieve faster and cheaper inference. Responsibilities include building performance models, analyzing workloads, enhancing tooling for bottleneck identification, and partnering with teams to implement improvements.

Skills and focus areas

• performance optimization
• benchmarking
• systems profiling
• distributed systems
• model inference
• hardware efficiency
• kernels
• accelerators
• networking
• fleet scheduling
• latency
• throughput