Senior TPU Software Engineer for ML Performance

Google; Sunnyvale (CA); On-site; USD 207,000 - 300,000; Full time

Benefits offered by this job

Equity

Bonus target 20%

Company benefits

Job summary

Google is seeking a software engineer for the Core ML team to optimize Gemini and OSS ML models on TPU systems. You will work across JAX and PyTorch platforms to boost performance for large-scale ML workloads.

The role emphasizes ML design, infrastructure, and collaboration with researchers, engineering, and product teams to push architectural limits of TPU hardware and software stacks.

Qualifications

• Bachelor’s degree or equivalent practical experience.
• 8 years of software development experience.
• 5 years in ML design/infrastructure and related areas.
• 3 years of software design/architecture experience.

Responsibilities

• Identify and maintain ML training and serving benchmarks for Google production.
• Achieve high performance for customer launches and OSS models.
• Drive out-of-the-box performance improvements for compiler/runtime.
• Collaborate with product teams to onboard new ML models on TPU hardware.
• Analyze performance metrics and implement fleet-wide improvements.

Skills

ML design and infrastructure

GPU accelerator expertise

System performance optimization

Education

Bachelor’s degree or equivalent practical experience

Master’s degree or PhD in Engineering/CS or related field

Tools

OpenXLA

MLIR

Triton

CUDA/OpenCL

Job description

Google is seeking a software engineer for the Core ML team to optimize Gemini and OSS ML models on TPU systems. You will work across JAX and PyTorch platforms to boost performance for large-scale ML workloads.

The role emphasizes ML design, infrastructure, and collaboration with researchers, engineering, and product teams to push architectural limits of TPU hardware and software stacks.