Join a pioneering team at Google Cloud, driving the future of machine learning infrastructure and Tensor Processing Unit (TPU) performance. You'll contribute to cutting-edge technologies that power large language models and empower cloud customers. This role offers an opportunity to innovate and shape the next generation of AI hardware and software stacks.

Our team is dedicated to advancing the machine learning software and hardware ecosystem, with a specific focus on the TPU compiler. We serve both internal Google teams and external Google Cloud Platform (GCP) clients, ensuring they have access to the most powerful ML tools available.

Design and implement advanced compiler optimizations, such as pipelining and fusions, to maximize TPU efficiency. Drive the performance of Tensor Processing Units (TPUs) for machine learning workloads. Collaborate across teams to integrate frameworks like PyTorch and ensure our ML platform remains industry-leading. You will gain an intuitive understanding of Google's ML stack, including frameworks like JAX and PyTorch, and the XLA runtime.

Research and develop novel compiler optimizations tailored for ML workloads and emerging architectures. Identify opportunities to enhance ML workload efficiency through insightful performance debugging for custom kernels. Provide technical leadership and mentorship, exploring strategic initiatives.

A Bachelor's degree or equivalent practical experience is required, complemented by 8 years of software development experience. You should possess 3 years of experience in software design and architecture. Essential technical skills include machine learning, compilers, computer architecture, GPU programming, C++, and Python. Experience with open-source development and ML compiler internals, including writing optimization passes, is highly preferred. Familiarity with debugging performance and correctness issues across the ML software stack and comfort with accelerator hardware architectures like TPUs and GPUs are also beneficial.