ML & Agentic Systems Engineer [IC4]
Sourcegraph1 day ago
Base Salary
$88k - $176k/yr
Responsibilities
- Design and harden reliable, observable, and cost-bounded multi-step agent systems for enterprise-scale products.
- Own model selection, upgrades, fine-tuning, distillation, evaluation, production rollout, and monitoring.
- Build representative datasets, evaluation harnesses, metrics, dashboards, error taxonomies, and release guardrails.
- Improve retrieval, ranking, context windows, citations, answer quality, cost, and latency.
- Drive technical direction for agent engineering and model-related roadmap decisions.
- Own high-risk technical projects end-to-end from problem framing through rollout and measurement.
- Mentor engineers through pairing, design reviews, code reviews, and agent-engineering guidance.
- Work with customers and translate feedback into product requirements, scopes, and milestones.
- Participate in the team’s on-call support rotation.
Requirements
- Staff-level technical leadership and experience owning ambiguous, high-impact engineering problems.
- Production experience owning a model lifecycle from dataset construction through evaluation, rollout, and monitoring.
- Experience designing multi-step agentic systems that are reliable, observable, and cost-bounded.
- Strong judgment in evaluation design, including datasets, baselines, metrics, error analysis, and release criteria.
- Ability to make quality, latency, and cost tradeoffs using model selection, prompting, retrieval, caching, distillation, and fine-tuning.
- Strong software engineering skills and ability to ship production services.
- Comfort with Go, TypeScript, GraphQL, Postgres, and Docker, or demonstrated ability to learn them.
- Autonomous product and technical problem-solving skills, including prototyping and end-to-end project ownership.
- Experience mentoring engineers and influencing technical direction beyond an immediate team.
- Customer-facing communication skills and ability to translate customer needs into product decisions.
- Preferred: experience shipping an LLM-powered or agentic developer product.
- Preferred: experience with retrieval, ranking, embeddings, search relevance, or enterprise customers.
Benefits
- Globally distributed, remote work environment with preferred residence in Europe or North America and at least 20 hours per week overlapping EST.
- Meaningful equity and generous perks and benefits.
- Base salary determined by the IC4 pay band and location zone, with transparent compensation bands.
- Interview process expected to take approximately 4.75 hours, including recruiter, technical, cross-functional, leadership, reference, and background-check stages.
Tech Stack
Categories
About Sourcegraph
Sourcegraph builds a code intelligence and AI context platform that provides code search, deep code understanding, and coordinated code changes across large, multi-repo codebases. It sells enterprise subscriptions to products including Code Search, Deep Search, MCP, and Agentic Batch Changes for engineering teams and AI agents. Privately held and founded in 2013, Sourcegraph is headquartered in San Francisco and counts companies like Stripe, Uber, Reddit, and Dropbox as customers.