ML & Agentic Systems Engineer [IC4]

Sourcegraph is seeking a Staff ML & Agentic Systems Engineer to join our Code Understanding team. We are building the infrastructure that allows engineering teams and AI agents to navigate and evolve complex codebases with confidence, and we are looking for a technical leader to drive our vision for agentic development.

Key responsibilities

• Design and harden multi-step, tool-using agent loops to create reliable, observable, and scalable products.
• Develop robust evaluation frameworks to measure agent performance, ensuring quality and cost-effectiveness.
• Manage the full lifecycle of models, including selection, upgrading, and fine-tuning for specific use cases.
• Optimize retrieval and context engineering to ground models accurately in customer codebases.
• Balance cost and latency as core product features through profiling, caching, and model distillation.

Requirements

• Proven experience owning a production model lifecycle, from dataset construction to rollout and monitoring.
• Fluency in building reliable, cost-bounded multi-step agentic systems.
• Strong judgment in evaluation, with the ability to build representative datasets and meaningful baselines.
• Ability to operate autonomously on ambiguous, high-risk technical problems.
• Strong software engineering fundamentals with experience in Go, TypeScript, and distributed systems.
• A pragmatic approach to shipping, prioritizing simplicity and high-quality standards.

What we offer

• The opportunity to shape the technical direction of agentic systems at a high-growth company.
• A globally distributed, remote-first team that values direct communication and high agency.
• Meaningful equity in a company backed by leading venture capital firms.
• A culture that prioritizes professional growth, mentorship, and customer-driven product development.