AMD - San Jose, CA
posted 5 months ago
At AMD, we are committed to transforming lives through our technology, and we are looking for a skilled and experienced engineer to join our core team. This role involves developing a cutting-edge machine learning model compiler specifically targeting AMD Inference Accelerator (AIE) hardware devices. The compiler will be responsible for taking models written in popular frameworks such as PyTorch, TensorFlow, ONNX, or JAX and producing optimized control and executable code for the AIE VLIW processor array. Additionally, the compiler must efficiently handle the partitioning of tasks between x86 and hardware accelerators, generate optimal code for AMD x86 CPUs using AMD ZenDNN, and create code that interfaces with the AMD AIE-specific runtime and driver. The ideal candidate will be passionate about implementing and improving effective algorithms and techniques to enhance compiler and runtime for machine learning. This position requires strong leadership skills to drive complex issues to resolution and the ability to communicate effectively while collaborating with various teams across AMD. The role also involves mentoring and providing guidance to others, learning the latest industry trends, and bringing innovative ideas to the team. You will be responsible for designing and developing groundbreaking AMD technologies, debugging and fixing existing issues, and researching alternative, more efficient methods to accomplish tasks. Building technical relationships with peers and partners is also a key aspect of this role.