Agent Engineer

Responsibilities

• Design and harden multi-step, tool-using agent loops for enterprise-scale products
• Implement pragmatic evaluation systems, including smoke tests and metrics, to measure agent performance
• Manage model selection, upgrades, and fine-tuning to optimize for quality and cost
• Drive retrieval and context engineering to improve model grounding and accuracy
• Optimize for cost and latency by profiling, distilling, and caching models
• Mentor teammates and set technical standards for agent engineering

Requirements

• Senior-level experience in agent engineering (software engineering, machine learning, and statistics)
• Proven track record of building, training, and operating models in production
• Experience designing reliable, multi-step agentic systems
• Strong software engineering skills with ability to ship production services
• Proficiency with Go, TypeScript, GraphQL, Postgres, or Docker
• Ability to work autonomously on ambiguous, high-risk technical problems
• Experience in an async-first, remote environment

Preferred Qualifications

• Experience shipping LLM-powered or agentic developer-facing products
• Experience building evaluation systems for LLM or ML products
• Expertise in fine-tuning, distilling, or training models for production targets
• Experience with retrieval, ranking, embeddings, or search relevance
• Experience working directly with enterprise customers

About the Company

Sourcegraph is a code intelligence platform that helps developers navigate and operate on massive, complex codebases. Backed by a16z and Sequoia, Sourcegraph provides tools like Code Search and Deep Search to empower millions of developers worldwide.