Description
Google is hiring a Software Engineer, TPU Infrastructure to bridge machine-learning workloads and custom Tensor Processing Unit hardware. The role focuses on benchmarking, workload characterization, distributed-system and compiler optimization, hardware-software co-design, performance modeling, and TPU mapping for production models such as LLMs and embedding models. The engineer will partner with Google Cloud, YouTube, Ads, and external customers to onboard workloads, improve performance, cost, and energy efficiency, and present hardware trade-offs to leadership. The position requires a bachelor’s degree or equivalent practical experience, two years of software development experience, two years of performance or large-scale-systems experience, and two years of computer-architecture or performance-analysis experience; compensation is $147,000–$210,000 USD plus a 15% bonus target, equity, and benefits.
