CUDA / GPU Optimization Engineer REMOTE — $80-$100/hr

About the role

Mercor is seeking elite GPU kernel optimization experts to collaborate on advanced projects with a leading AI lab. If you enjoy squeezing every drop of performance out of modern GPU architectures and want to work at the frontier of AI infrastructure, this is for you.

Responsibilities

• Analyze and fiercely optimize GPU kernels for maximum performance, efficiency, and hardware utilization.
• Leverage profiler metrics (L2 cache hit rate, occupancy, etc.) to guide kernel improvements.
• Write, modify, and reason about complex C++17, Python, and GPU programming code.
• Document optimization decisions clearly and efficiently.

Requirements

• Languages: Fluent in core C++ (through C++17) and Python.
• GPU Ecosystem: Fluent in at least one major GPU programming model (CUDA, HIP, Slang, HLSL, GLSL).
• Experience: 1+ years of professional or graduate-level research experience directly working with GPUs.
• Deep Understanding: Strong grasp of GPU profiler metrics (NSight Compute familiarity is a big plus).
• Bonus: Prior experience with NVIDIA (especially Blackwell), AMD, or Qualcomm.

Pay

Rate: $80 - $100 USD per hour.

Schedule

Minimum 20 hours/week commitment (Flexible, 100% Remote schedule).

Visa Notice for US Applicants: This is an independent contractor role. You must be legally authorized to work as a contractor. We are strictly unable to support H1-B or STEM OPT candidates for this project.