Sourcegraph
Agent Engineer [IC4]
About this role
Join Sourcegraph as a senior Agent Engineer to lead the technical direction of AI-powered code understanding features. You'll own the design and optimization of multi-step agentic systems, model selection and training, retrieval strategies, and cost/latency tradeoffs—turning research into reliable, enterprise-scale products that help developers navigate massive codebases.
What you'll do
- Design and harden multi-step, tool-using agent loops for agentic code understanding experiences
- Establish evaluation frameworks and metrics to measure agent quality while moving fast
- Select, upgrade, and fine-tune language models for different use cases and cost requirements
- Optimize retrieval, ranking, and context engineering to ground models in customer code
- Profile and optimize for cost and latency as core product features
- Set technical standards and raise team fluency in building with AI models
What they're looking for
- Agent design and multi-step LLM systems
- Machine learning and LLM fine-tuning
- Evaluation design and metrics for agentic systems
- Retrieval and ranking systems
- Cost and latency optimization
- Software engineering and systems design
- Technical leadership and mentoring
- Working with code understanding and search systems
Benefits
- Remote work with flexible location (preference for Europe/North America)
- Work on infrastructure used by developers at major companies like Stripe and Uber
- Backed by top-tier VCs (a16z, Sequoia, Redpoint)
- Direct impact on product direction and customer outcomes
- Senior role with real agency over technical decisions
- Small, senior-leaning team with streamlined product ownership
Opens the official application on the employer’s site. No login required.
Sourcegraph
Sourcegraph builds a code intelligence platform that helps developers understand and navigate large codebases through AI-powered features and open-source indexing protocols. The company is hiring for infrastructure engineers, AI/agent specialists, compiler engineers, security professionals, and customer-facing field engineers to scale its SaaS offering and enterprise deployments.
- Website
- sourcegraph.com
Likely interview questions
- Walk us through how you'd approach designing a multi-step agent system for code understanding—what are the key decisions you'd make around model selection, retrieval, and tool use?
- How do you think about evaluating agent systems when the product and models are constantly changing? What tradeoffs do you make between rigor and speed?