Description
Amazon is hiring a Kernel Engineer for its Neuron Compiler organization to design and implement high-performance compute kernels for machine learning operations on Inferentia and Trainium accelerators. The role involves optimizing kernel-level performance, using profiling and compiler techniques such as fusion, sharding, tiling, and scheduling, collaborating with customers on model enablement, and mentoring engineers. It requires at least five years of professional software development, programming, system architecture, and engineering leadership experience, with accelerator, GPU, and low-level optimization expertise preferred. The position is based in Toronto, Canada, with a base salary of CAD 150,700 to 251,700 annually.
