Unknown Company

Remote CUDA Kernel Performance Engineer

workfromhome • Posted 2 days ago
Remote Contract Computer and Mathematical Occupations

Mercor is seeking a CUDA Engineering Expert to analyze and optimize GPU kernels for performance, leveraging CUDA, HIP, and shader programming to improve efficiency in AI research workloads. This remote contract role requires fluent C++17, Python, and Git, and at least 1 year of GPU-focused experience.

Candidates will document optimization choices, review kernel implementations, and work 20 hours per week. Prior experience with NSight Compute and NVIDIA/other hardware is a plus.

#J-18808-Ljbffr

Remote CUDA Kernel Performance Engineer in workfromhome at Unknown Company

This position is listed as contract and able to be worked remotely.

Back to Job Search