Research Engineer - CUDA Kernel Engineering

What you'll do

• Develop, integrate, and optimize CUDA kernels for AI and hardware-design workloads.
• Build performance benchmarks and integration layers to improve GPU utilization across large-scale systems.
• Collaborate with AI researchers and semiconductor experts, and contribute kernels and tooling to open-source AI and HPC ecosystems.

Nice to have

• Experience writing and optimizing CUDA kernels for large-scale AI workloads.
• Experience profiling GPU performance, integrating custom kernels with training and inference frameworks, or working with NVIDIA hardware and software stacks.

Tools

CUDA, PyTorch, Megatron, vLLM, TorchTitan, Hopper, Blackwell, NVLink

AI focus

Build GPU kernels and tooling for AI models used in semiconductor design and verification, including circuit-layout reasoning, RTL generation and validation, chip-architecture optimization, training, inference, and reinforcement learning.