Senior RAG Engineer

Own retrieval for our agent platform — hybrid search, reranking, and an eval harness that gates every release.

• Design hybrid retrieval (BM25 + dense + rerank)
• Build offline + online retrieval evaluation
• Drive p95 query latency under 200ms at scale
• Mentor two mid-level engineers

• 3+ yrs shipping production RAG or search
• Strong Python; comfort in a typed language
• Verified contribution history (Inferr profile)

Halyard builds autonomous research agents used by thousands of scientists. Small team, deep technical bar, remote since day one.