Principal / Senior GPU SW Performance Engineer — Post‑Training - CA, San Jose, United States

Originally posted 25 March 2026 by the employer. Relisted by the employer on 6 May 2026.

About the role

Drive the performance of post-training workloads on AMD Instinct GPUs across kernels, distributed training, and framework integrations to deliver training pipelines on ROCm.

What you'll do

- Lead performance for finetuning and RL training solutions on AMD GPUs. - Improve throughput, memory efficiency, and stability across data, model, and optimizer steps. - Optimize multi-GPU/multi-node training and communication patterns. - Contribute efficient kernels/ops and targeted graph-level optimizations. - Profile, diagnose, and resolve bottlenecks using standard tooling, and prevent regressions in CI. - Ship reproducible pipelines and documentation adopted by internal teams and external developers.

What you'll need

- B.S./M.S./Ph.D. in Computer Science, Computer Engineering, Electrical Engineering, or equivalent.

Nice to have

- Proven GPU performance engineering for deep learning (ROCm/HIP, Triton, or similar). - Hands-on experience with SFT, LoRA, and RL-based training at scale. - Strong PyTorch experience (torch.distributed, FSDP/ZeRO or equivalent). - Proficiency in Python and C++, with comfort reading/writing kernels when needed. - Experience with distributed systems and collective communication libraries. - Track record of turning profiles into fixes, upstreaming changes, and documenting results.

Skills: GPU performance engineering, deep learning, ROCm/HIP, PyTorch, distributed systems