About the job
Responsibilities
Develop the PyTorch-Native Kernel Programming Model by designing an integration layer and runtime environments that facilitate the execution of custom kernels as native extensions in the PyTorch framework.
Create a Tensor-Level Kernel Language by establishing high-level abstractions and programming models to support efficient model integration and expressive tensor operations.
Implement a Virtual ISA to address performance bottlenecks through direct hardware control while preserving high-level programmability.
Contribute to the Kernel Developer Ecosystem by building essential programming tools, technical specifications, and reference implementations to streamline kernel development and foster community contributions.

