New: upload your resume and get matched to AI jobs. Free to start →

Software Engineer III, Machine Learning Compiler, TPU Scaling

$147k - $210k • Sunnyvale, CA, USA • Full-time

Posted 8d ago

Remote Work Policy

On-site

Employment Type

Full-time

Categories

Machine Learning Engineer

About the job

Our team develops the ML Compiler (XLA) for running workloads on thousands of TPU, GPU or other accelerators. Accelerated Linear Algebra (XLA) powers all Google’s internal ML workloads and is a preferred choice for external foundational model producers and researchers for its advanced performance, usability and open source nature. As part of the team, you will be developing the compiler's ability to generate efficient communication on TPU. Your team will develop the compiler's horizontal scaling stack including collective op lowering, hiding communication latency, program sharding and more. Together these make it possible to train and serve the world's largest models efficiently on TPU.

Responsibilities

  • Execute solutions in specialized ML areas, utilize ML infrastructure, and contribute to model optimization and data processing.
  • Implement and optimize collective operations in the XLA TPU backend, from HLO-level transformations to low-level code generation.
  • Improve compiler scheduling to overlap communication with computation, balancing latency, memory pressure, and throughput.
  • Analyze current and upcoming TPU architecture and interconnect features to improve scaling performance.
  • Work with ML performance and research teams to diagnose scaling bottlenecks and achieve roofline performance on critical workloads.

Requirements

  • Bachelor's degree or equivalent practical experience.
  • 2 years of experience in low-level systems programming languages (e.g., C++ or C).
  • 1 year of experience with Speech/audio, reinforcement learning, ML infrastructure, or another ML specialization.
  • 1 year of experience with end-to-end Machine Learning (e.g., model deployment, evaluation, data processing, debugging).
  • Experience in performance optimization and compiler development.
  • Master's degree or PhD in Computer Science or related technical fields (preferred).
  • 2 years of experience with data structures and algorithms (preferred).
  • Experience developing accessible technologies (preferred).
  • Experience with Hardware Design, Hardware Architecture (preferred).

Benefits

  • 15% bonus target
  • Equity
  • Benefits

About Google