CUDA Kernel Engineer

About Periodic Labs

We are an AI + physical sciences lab building state of the art models to make novel scientific discoveries. We are well funded and growing rapidly. Team members are owners who identity and solve problems without boundaries or bureaucracy. We eagerly learn new tools and new science to push forward our mission.

About the role

You will develop, integrate and optimize state-of-the art CUDA kernels to power AI scientific research. You will integrate CUDA kernels into training, inference and reinforcement learning systems running on thousands of GPUs. You will build tools and directly support frontier-scale experiments to make Periodic Labs the world’s best AI + science lab. You will release your kernels as contributions to the open-source AI stack.

You might thrive in this role if you have experience with:

• Writing and optimizing CUDA kernels: attention, mixture-of-experts, dispatch-and-combine, and others

• Working with the latest generation of Nvidia hardware

• Integrating kernels into state-of-the-art inference (vLLM, SGLang) and training frameworks (Megatron, TorchTitan)