Software Engineer, Workload Enablement

The role

This role enables production workloads and end-to-end testing on new platforms. You will create test harnesses, port workloads to new hardware, and analyze performance and bottlenecks.

What you will do

• Port and validate key inference and training workloads on new platforms/SKUs
• Build a suite of benchmarks and stress tests for system aspects including CPU, GPU, memory, networking, and storage
• Deep-dive performance on distributed training/inference
• Perform collective performance and tuning across NCCL/RCCL and internal libraries
• Create repeatable test harnesses that run in CI/lab environments
• Partner with systems and fleet bring-up engineers for platform stability and scalability
• Work cross-functionally with vendors and internal stakeholders to produce bug reports and issue lists

Required qualifications

• BS in CS/EE or equivalent practical experience
• 5+ years in ML systems, performance engineering, distributed systems, or HPC
• Hands-on experience with PyTorch and modern LLM training/inference stacks
• Hands-on experience with large-scale distributed training concepts
• Experience with RDMA and debugging/optimizing comms libraries (NCCL or RCCL)
• Proficiency in Python
• Comfort reading/writing performance-critical code
• Strong profiling/debugging skills

Nice to have

• C++/CUDA/HIP
• Experience building workload-shaped benchmarks and stress/fault tests
• Familiarity with RDMA networking and transport tuning
• Experience running and validating workloads in Kubernetes
• Hands-on lab experience with early hardware

About the company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity.