Performance Engineer, GPU

Overview

As a GPU Performance Engineer, you architect and implement foundational GPU performance and systems capabilities that power Claude and improve large language model inference efficiency.

What you'll do

• Maximize GPU utilization and performance at scale for production ML systems.
• Develop optimizations across the stack from low-level tensor core work to system orchestration.
• Implement GPU kernel and optimization techniques, including kernel fusion and memory bandwidth optimization.
• Design and optimize distributed training/inference communication and multi-node GPU synchronization.
• Build performance modeling frameworks to predict and optimize GPU utilization.
• Profile and eliminate performance bottlenecks in production serving and training infrastructure.

What you'll need

• Deep experience with GPU programming and optimization at scale.
• Experience shaping measurable GPU performance breakthroughs.
• Ability to navigate system complexity from hardware interfaces to high-level ML frameworks.
• Experience with collaborative problem-solving and pair programming.
• Excitement to work on state-of-the-art language models with real-world impact.
• Comfort working in ambiguous environments where you define the path forward.

Details

• Location-based hybrid policy: staff are expected to be in an office at least 25% of the time.