Description
Google DeepMind is hiring a Software Engineer to improve the training and serving efficiency of Gemini models on TPUs and GPUs. The role focuses on profiling distributed workloads, designing low-level custom kernels, building automated benchmarking and diagnostic tooling, and influencing accelerator architectures and compiler roadmaps. Candidates need a bachelor’s degree or equivalent practical experience, at least two years of C++ or Python software development experience, profiling-tool experience, and experience developing software for hardware accelerators.
