Software Engineer, Search Systems - Code Data
Core
Design and build end-to-end architecture and algorithms for searching across code tasks, routing them to the right models, and translating natural-language questions into precise retrieval.
Role type
Staff-level software engineer (search & retrieval systems)
Builds
Code search and retrieval systems that combine dense embeddings and lexical signals for frontier coding models
Domain
AI data, code search, retrieval systems
Deliverable
production ML models
Required skills
hybrid retrieval (dense embeddings + BM25), candidate generation, ranking, re-ranking, ANN indices (HNSW, IVF), inverted indices, Elasticsearch/OpenSearch/Lucene/FAISS, distributed systems, data modeling, API design, cost-vs-speed tradeoffs, query understanding, technical leadership, mentorship
Preferred skills
code search or code understanding, code embeddings, code-specific models, semantic parsing, LLM-assisted query generation
Technologies
HNSW, IVF, product quantization, Elasticsearch, OpenSearch, Lucene, FAISS, vector databases
Responsibilities
Own architecture of code search and retrieval systems; solve hard problem of identifying similar code tasks; build system to select code-specific models; design natural-language-to-query translation; design and operate indexing pipeline; make cost-and-speed tradeoffs; build evaluation harnesses; define long-term technical strategy; establish evaluation metrics and quality guardrails; mentor engineers; partner on build-vs-buy decisions
Seniority
Staff, hands-on technical leadership
