Agent Engineer

Hours & location 🌎 While we hire almost anywhere in the world, we have a preference for someone to reside in the following locations for this role. However, if you feel qualified, we welcome you to apply regardless of location. No matter what, working hours must overlap with EST for at least 20 hours/week. Preferred locations: Europe North America Why this job is exciting Sourcegraph is at the forefront of building AI tools to solve the biggest problems in the software industry, problems that only get bigger as codebases grow and as more of the work is done by agents. The Code Understanding team owns the surfaces where that intelligence meets the developer: Deep Search, our agentic, multi-step answer engine across an enterprise's entire lineup of codebases, Query Assist, turning natural language queries Sourcegraph query syntax, Smart Hovers, concisely summarizing symbols right where devs need it, guided diff review, and the APIs that both humans and AI agents rely on every day. Most of what makes those surfaces tick is agent engineering: a blend of software engineering, machine learning, and statistics. Agent engineering tells us which model to use and when, how to retrieve and pack context, how to measure answer quality, where to fine-tune or distill a smaller model to cut costs and latency, and how to expand a single LLM call into a reliable multi-step agent. As the senior agent engineer on Code Understanding, you'll be the technical owner of it: setting the agentic direction for the team, making our products measurably better, faster, and cheaper, and raising the team's fluency in building with models. This is a senior role: we're hiring a technical leader, not just a strong individual contributor. You'll own the hardest, most ambiguous problems in this space, set standards others follow, and influence direction beyond your immediate team. You'll get the exhilarating chance to drive the vision on how we can provide the best code understanding experience on the market by combining our deterministic, large-scale systems and AI into experiences never seen before. Concretely, you'll own work like: Agentic systems. You'll design and harden the multi-step, tool-using agent loops behind current and new agentic experiences, turning research and experiments into reliable, observable, and affordable products at enterprise scale. Pragmatic use of evaluations. Crafting agentic products means needing to tell when a change actually helped, which is hard when agents keep changing and the product keeps shifting. You'll bring judgment about where evaluations earn their keep, when to use targeted smoke tests and metrics, and how to avoid noise dressed up as rigor, so we can move fast with confidence. Models: selection, upgrading, and training. You'll decide which models we run where, drive upgrades, and fine-tune our own when that's the right call. Retrieval and context engineering. You'll push on how we ground models in a customer's code - retrieval, ranking, context windows, citations - to make answers more accurate and verifiable. Cost and latency. Every surface has a per-user economic budget. You'll treat cost and latency as product features, and profile, distill, cache, and right-size models so we can ship ambitious features sustainably. You'll do this on a small, senior-leaning team that ships quickly, owns a lot of product surface, and has streamlined product management: engineers here talk to customers, frame the problem, and own it end-to-end. You'll have real agency over technical direction and a direct line to the impact of your work.