Performance SOC Architect - GPU/ML Profiling & Optimization

Google LLC; Sunnyvale (CA); On-site; USD 174,000 - 252,000; Full time

Job summary

Google LLC in Sunnyvale, CA is seeking a Senior Software Engineer focused on Performance and SOC Architecture. You will drive performance analysis, GPU- and accelerator-focused optimization, and data-driven debugging for large-scale systems.

The role emphasizes collaboration across Hardware, Runtime, and ML teams, with a strong emphasis on profiling workloads, kernels, and system performance to improve developer productivity.

Qualifications

• Bachelor’s degree or equivalent practical experience.
• 5 years of experience with software development in one or more programming languages.
• 3 years of experience with performance, large-scale systems data analysis, visualization tools, or debugging.
• 3 years of experience with Computer Architecture, C++, Python, CUDA, GPU Programming, GPU Drivers, System On a Chip, ASIC.

Responsibilities

• Write and test product or system development code.
• Learn and build an intuitive understanding of existing data collection, analysis, and visualization workflows.
• Support new and exciting ML paradigms by making contributions across the end-to-end stack and analysis tools.
• Partner with Product Area leads to understand model optimization use cases and chip profiling requirements.
• Collaborate across Hardware, Driver, Runtime, and Performance Analysis teams and stakeholders.

Skills

Software development

Performance analysis

Data visualization

Debugging

Computer architecture

C++

Python

Education

Bachelor’s degree or equivalent

Tools

CUDA

GPU programming

GPU drivers

SoC

ASIC

TensorFlow

JAX

PyTorch

Job description

Google LLC in Sunnyvale, CA is seeking a Senior Software Engineer focused on Performance and SOC Architecture. You will drive performance analysis, GPU- and accelerator-focused optimization, and data-driven debugging for large-scale systems.

The role emphasizes collaboration across Hardware, Runtime, and ML teams, with a strong emphasis on profiling workloads, kernels, and system performance to improve developer productivity.