Model Performance Software Engineer, Claude Code
Core
Set technical direction for evaluation systems, research infrastructure, and internal tooling to measure, understand, and improve Claude's coding capabilities.
Role type
Staff Software Engineer (Technical Lead)
Builds
Evaluation frameworks, research infrastructure, and internal tooling for the Claude Code team
Domain
AI Safety / Large Language Models / Software Engineering
Deliverable
infrastructure
Required skills
Architecting complex high-stakes systems, setting technical direction, mentoring engineers, translating research questions to engineering solutions, owning critical systems end-to-end
Preferred skills
Designing eval/evaluation frameworks for ML systems, reinforcement learning infrastructure, research computing at scale, Python, TypeScript
Technologies
Python, TypeScript
Responsibilities
Set technical direction for evaluation systems and research infrastructure; Architect eval frameworks measuring model capabilities; Lead design of infrastructure enabling large-scale experiments; Identify high-leverage engineering investments; Serve as technical bridge between product and research; Mentor engineers and review designs; Partner with research leads to solve ambiguous problems; Own critical systems from architecture to production reliability
Seniority
Staff, hands-on IC with technical leadership
