Senior Software Engineer, TorchTPU

Overview

In this role, you will advance AI framework development to run PyTorch on Google Cloud TPUs and GPUs, optimizing performance at scale. You will work within the Core ML/PyTorch TPU team to support ML frameworks and compilers across TPU pods and cloud GPUs. You’ll enable large-scale generative models, CV, NLP, and other workloads, impacting how customers train and deploy state-of-the-art models in the cloud. This is an opportunity to shape PyTorch-in-Cloud and contribute to open-source efforts with cross-functional partners.

Responsibilities

• Develop and optimize PyTorch on TPUs/GPUs for cloud training and inference
• Provide comprehensive ML framework and compiler support in Cloud TPU/GPUs environments
• Scale and tune models across domains (CV, NLP, speech, recommendations) for performance and cost
• Collaborate with Google teams and researchers to advance ML capabilities in PyTorch-in-Cloud
• Contribute to open-source software development and ecosystem engagements

Key requirements

• Bachelor’s degree in Computer Science or related field; equivalent practical experience
• 5+ years of software development experience (C++, C, or Python)
• Experience leading cross-functional software projects from inception to delivery

• effective communication in fast-paced environments
• cross-functional collaboration
• leadership and ownership mindset
• Experience with ML frameworks (TensorFlow, JAX, PyTorch)
• Hardware performance analysis on TPUs and GPUs
• Compiler experience / compiler construction

…