Teams at Motional

Develop and optimize the systems powering large-scale distributed machine learning training. Responsibilities include profiling performance with Nsight and PyTorch Profiler, optimizing data loading and communication, developing GPU kernels in Triton or CUDA, and improving training throughput, reliability, and cost using PyTorch Distributed.

As a Machine Learning Systems Engineer, you will optimize distributed training systems, enhance performance profiling, and develop high-performance GPUs for machine learning workloads.

Develop and optimize systems for large-scale distributed machine learning training. Profile and improve data loading, gradient computation, communication, and model execution using tools such as Nsight and PyTorch Profiler. Build high-performance GPU kernels with Triton or CUDA, optimize PyTorch Distributed pipelines, and improve training throughput, reliability, cost, and time-to-convergence.