Performance Engineer, GPU

GPU Performance Engineer role involves architecting and implementing foundational GPU performance systems to maximize utilization and inference efficiency for large language models. Responsibilities span from low-level tensor core optimizations to coordinating thousands of GPUs in distributed environments, with opportunities to develop custom kernels, co-design attention mechanisms, and optimize end-to-end training and inference pipelines. Preferred skills include CUDA/Triton/CUTLASS, Flash Attention, tensor core optimization, NCCL/NVLink, mixed-precision, and experience with production-scale ML infrastructure. Visa sponsorship is offered; hybrid in-office policy expects presence at least 25% of the time; location shown is San Francisco, CA, USA.