GPU Performance Engineer: Scale AI Inference

Anthropic; San Francisco (CA); On-site; USD 315,000 - 560,000; Full time

Benefits offered by this job

Competitive salary

Equity opportunities

Flexible working hours

Generous vacation and parental leave

Job summary

A leading AI research company in San Francisco is seeking a mid-senior GPU Performance Engineer. In this role, you'll architect systems that enhance GPU performance for groundbreaking AI models. Responsibilities include developing optimizations, collaborating with teams, and troubleshooting performance issues. The position offers a competitive salary range of $315,000—$560,000, equity opportunities, and a supportive work environment that values communication and collaboration.

Qualifications

• Deep experience with GPU programming and optimization at scale.
• Able to navigate complex systems from hardware to high-level frameworks.
• Passionate about delivering measurable performance breakthroughs.

Responsibilities

• Architect and implement systems for GPU performance in AI.
• Develop and optimize end-to-end training pipelines.
• Collaborate with researchers for impact-driven solutions.

Skills

GPU programming and optimization at scale

Collaborative problem-solving

Navigating complex systems

Impact-driven mindset

Education

Bachelor's degree in a related field or equivalent experience

Tools

CUDA

PyTorch/JAX internals

NCCL

Nsight

Job description

A leading AI research company in San Francisco is seeking a mid-senior GPU Performance Engineer. In this role, you'll architect systems that enhance GPU performance for groundbreaking AI models. Responsibilities include developing optimizations, collaborating with teams, and troubleshooting performance issues. The position offers a competitive salary range of $315,000—$560,000, equity opportunities, and a supportive work environment that values communication and collaboration.