Performance Engineer, Inference Engine

Highlights

• Optimize in-house inference engine serving Claude to millions of users.
• Improve throughput, cost, reliability and latency across accelerator platforms.
• Drive measurement loop that makes Claude faster with every iteration.

What they offer

• Competitive compensation and benefits
• Optional equity donation matching
• Generous vacation and parental leave
• Flexible working hours
• Lovely office space for collaboration

Ideal candidate

A systems-minded performance thinker who ramps fast in deep unfamiliar codebases, profiles before changing code, and loves pairing. Low-ego collaborator motivated by making Claude faster, reliable, and safe for society.

Requirements

• Working mental model of LLM inference across compute, memory, and interconnect
• Proven quick learner who ramps fast and ships consequential changes
• Strong systems programming in Rust, C++, or similar with care for quality
• Analytical about performance: profile, hypothesize, test, then measure
• Low ego, takes feedback well, and picks up slack outside job description
• Enjoys pair programming and cares about societal impacts of work
• Bachelor's degree or equivalent combination of education and experience