Sourcegraph is the code-intelligence platform — universal code search across repositories, plus Cody (an AI coding assistant). Used by Uber, Lyft, Reddit, GE, and many others. The interview emphasizes code-graph engineering, large-scale indexing (Zoekt, SCIP), and the systems work behind LSP-like tooling at organization scale.
Process
Recruiter screen → 60-minute coding (DSA medium, language of choice) → onsite virtual: 2 coding, 1 system design, 1 craft deep-dive, 1 behavioral. Cycle: 3–4 weeks.
What they actually ask
- Design a code-search index over millions of repositories
- Design a precise code-graph using SCIP/LSIF for cross-repo navigation
- Design RAG over a code corpus for an AI coding assistant
- Coding: medium DSA, often with parsing, trie, or graph framing
- Behavioral: developer empathy, ownership, written-first culture
Levels and comp (2026)
- Engineer (E3): $175K–$230K total
- Senior Engineer (E4): $230K–$310K
- Staff (E5): $310K–$420K
- Principal (E6): $420K–$580K
(Bands published openly on the Sourcegraph handbook.)
Prep priorities
- Be fluent in Go (most of the backend) and TypeScript (frontend, Cody)
- Understand code-indexing concepts (Zoekt n-gram search, SCIP/LSIF graphs, tree-sitter)
- Brush up on RAG patterns over code (chunking, embeddings, code-aware retrieval)
Frequently Asked Questions
Is Sourcegraph remote-friendly?
Fully distributed. No physical HQ. Engineers across the Americas, Europe, and APAC. Annual all-hands in person.
How does Sourcegraph compare to GitHub Copilot or Cursor?
Copilot is editor-embedded for individual devs. Cursor is an AI-first IDE. Sourcegraph (Cody + Code Search) emphasizes large-org code-graph context. Comp is competitive at senior+ with strong remote flexibility.
What is the engineering culture?
Async, written-first (everything in handbook/RFCs), high autonomy, ship-focused. Strong fit for engineers who prefer writing over meetings.