Machine Learning Performance Engineer

What you'll do

• Improve performance of large-scale training, low-latency real-time inference, and high-throughput research inference.
• Use a whole-systems approach to optimize storage, networking, host, and GPU performance.
• Investigate low-level platform performance, including throughput, goodput, and memory access.

What they ask for

• Understanding of modern ML techniques and toolsets.
• Systems knowledge to debug training-run performance end to end.
• Low-level GPU knowledge, including PTX, SASS, warps, cooperative groups, Tensor Cores, and memory hierarchy.
• Experience debugging and optimizing with tools such as CUDA GDB, NSight Systems, and NSight Compute.

Nice to have

• Knowledge of GPU libraries including Triton, CUTLASS, CUB, Thrust, cuDNN, and cuBLAS.
• Background in GPU-cluster networking and distributed training with NCCL or MPI.

Tools

CUDA, PTX, SASS, Triton, CUTLASS, cuDNN, cuBLAS, NCCL

AI focus

Optimize machine-learning model training and inference, including distributed GPU training.