LLM & Retrieval Engineer

Role overview Lead retrieval and context engineering for production LLM systems, improving answer quality through deliberate design and measurement. The work spans retrieval pipelines, evaluation, prompt and context strategies, and the infrastructure that keeps those systems responsive and cost-aware.

Responsibilities - Design and tune retrieval pipelines, including chunking, embeddings, hybrid search, and reranking. - Create evaluation datasets and metrics to measure retrieval and generation quality. - Develop context and prompt strategies that handle edge cases reliably. - Track and manage retrieval cost and latency at production scale. - Partner with data engineering to build and maintain the pipelines that supply search indexes.

Requirements - At least six years of engineering experience, with recent, focused work on LLM and retrieval systems. - Hands-on experience with vector databases such as pgvector, Pinecone, or Weaviate, and with hybrid retrieval. - Ability to evaluate retrieval and generation both offline and in production. - Strong Python skills and familiarity with the data pipelines behind an index. - A measurement-led approach to improving system quality.

Nice to have - Experience with fine-tuning or distillation for cases where retrieval alone is insufficient. - Familiarity with evaluation tools such as Braintrust or Langfuse.

Benefits and work setup - Full-time, remote role for candidates in North America or EMEA. - Listed compensation is $180,000–$225,000 plus equity.