Google is a leading technology company that develops next-generation technologies to enhance user connectivity and interaction. They are seeking a Software Engineer to work on the EdgeTPU compiler team, focusing on optimizing and enhancing compiler performance for machine learning workloads.
Analyze and improve compiler quality, performance, optimization decisions, correctness and compilation time as part of the EdgeTPU compiler team
Develop and enhance parallelization and scheduling algorithms to reduce compute and data movement costs for executing ML workloads on the EdgeTPU
Collaborate with EdgeTPU architects to shape future accelerators, the Hardware/Software (HW/SW) interface, and co-optimizations of the next-generation EdgeTPU architectures
Work on efficient mapping of generative AI models and other key workloads into EdgeTPU instructions through the compiler
Qualification
Required
Bachelor's degree or equivalent practical experience
2 years of experience working with embedded systems or device software
2 years of experience with compilers (development, parallelization, etc.)
2 years of experience with software development in one or more programming languages (e.g. C++), or 1 year of experience with an advanced degree
Preferred
Master's degree or PhD in Computer Science or related technical fields
Experience in optimizing ML models for inference
Experience developing compilers within the context of accelerator-based architectures
