Job Details

The Culture We Are Building

• We are assembling a lean group of generalists. While we welcome experience, we're primarily looking for creative, curious problem solvers with a keen interest in open-ended, technically challenging problems.
• We are looking to hire in-person in San Francisco. We believe that having a community is critical for learning, growth, and camaraderie. We will pay generously for relocation assistance.
• We want to invest heavily in our team's growth. We will happily sponsor conference attendance and encourage projects and research. If there's something that will help you grow as an engineer, researcher, or designer, we actively want to support it.
• We believe in compensating exceptional talent exceptionally well. We are an early stage company, but are committed to working with candidates we're excited about to ensure they are excited too.
• We have an academic as well as competitive spirit. We enjoy thought exercises, math puzzles, intellectual inquiry, and vigorous debates.

Role: Systems Engineer

Build developer tooling for semantic diffing and code search. Ship fast search across codebases and diffs, a diff engine that understands syntax trees and symbols, and workflows including 'rebase a review.' Help invent new diffing primitives, recognize refactors and code movement, and advance the version-control model itself.

Responsibilities

• -Design and implement fast, accurate full-text and vector search over diffs and codebases; build embedding and indexing pipelines and benchmark retrieval quality, latency, and cost
• -Evaluate and integrate vector stores (e.g., TurboPuffer, S3 Vectors, pgvector, Chroma) or build bespoke indexers optimized for code and diffs
• -Build a semantic diff engine that understands syntax trees, symbols, and refactors; detect renames and code movement across files, functions, and modules
• -Design algorithms to recognize equivalence under renaming, indentation, and structural transformations; enable change grouping and conflict visualization
• -Deliver low-latency, high-throughput indexing and retrieval at massive-repo scale; define SLOs, instrument, and continuously benchmark
• -Prototype and productionize workflows like "rebase a review" and multi-VCS support (Git, Jujutsu), including deep integrations with code hosts

Qualifications

• -Expertise in Rust, C, or Zig and systems programming, with strong understanding of performance, concurrency, and correctness
• -Experience with ASTs/parsers (e.g., tree-sitter), program analysis, or compiler/tooling internals
• -Hands-on background in search and retrieval systems: embeddings, vector indexes, ranking, or RAG
• -Understanding of version-control internals (Git DAGs, merge/rebase strategies); exposure to alternative VCS (e.g., Jujutsu) is a plus
• -Experience building scalable indexing pipelines and services, with strong observability and benchmarking discipline