Software Engineer III, Performance

Google LLC; Mountain View (CA); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Bonus

Equity

Benefits

Job summary

Google LLC is seeking an engineer to optimize ML models for Google TPUs, employing sharding, quantization, and custom kernels to boost performance. You will build scalable performance tools and optimization recipes for customers, collaborate on compiler cost modeling, and maintain benchmarks to reveal performance trends.

The role requires a bachelor’s degree or equivalent experience, 2+ years in software development, and expertise in computer architecture, performance analysis, and modeling.

Qualifications

• Bachelor's degree or equivalent practical experience.
• 2+ years of software development experience.
• 2+ years of experience in computer architecture, performance analysis, and performance modeling.
• Experience with AI agents or LLM-powered workflows and systems engineering; preferred qualifications include accelerator programming, ML compilers, runtime systems, and accelerator architectures.

Responsibilities

• Optimize machine learning models for Google TPUs using techniques such as sharding, quantization, and custom kernels.
• Develop scalable performance tools and optimization recipes for customers.
• Collaborate on compiler cost modeling and maintain benchmarks and trend analysis to identify performance issues.

Skills

Software Development

Computer Architecture

Performance Analysis

Performance Modeling

Systems Engineering

AI Agents

Low-Level Accelerator Programming

CUDA

Triton

Machine Learning

Compiler Optimization

Code Generation

Runtime Systems

Distributed Systems

Performance Tuning

Education

Bachelor's degree or equivalent

Tools

CUDA

Triton

Pallas

Job description

Optimize machine learning models for Google TPUs using techniques such as sharding, quantization, and custom kernels, and develop principles that can be automated with agentic AI. Build scalable performance tools and optimization recipes for customers, collaborate on compiler cost modeling, and maintain benchmarks and trend analysis to identify performance issues.

Requirements: Requires a bachelor's degree or equivalent practical experience, at least two years of software development experience, and two years of experience in computer architecture, performance analysis, and performance modeling. Candidates also need experience with AI agents or LLM-powered workflows and systems engineering; preferred qualifications include advanced education and expertise in accelerator programming, ML compilers, runtime systems, and accelerator architectures.

Key Skills: Software Development, Computer Architecture, Performance Analysis, Performance Modeling, Systems Engineering, AI Agents, Low-Level Accelerator Programming, CUDA, Triton, Pallas, Machine Learning, Compiler Optimization, Code Generation, Runtime Systems, Distributed Systems, Performance Tuning

Benefits: Bonus, Equity, Benefits