Principal Applied AI Search & Hybrid Retrieval Engineer at LexiSearch

Job Description

LexiSearch delivers sub-second semantic and lexical search across hundreds of millions of regulatory filings, patents, and engineering specifications. We are hiring a Principal Applied AI Search and Hybrid Retrieval Engineer.

Role Overview You will train domain-specific bi-encoder embeddings and late-interaction ColBERT rerankers in PyTorch, tune HNSW vector indexes alongside BM25 lexical scoring, and maintain our evaluation harness measuring NDCG@10 and zero-hallucination citation recall.

Qualifications Seven+ years building information retrieval, search relevance, or applied NLP systems in Python. Hands-on experience with vector databases, pgvector/PostgreSQL, or custom inverted indexes. Strong experimental rigor with offline golden datasets and online interleaving A/B tests.