Own and advance a production physician-organization matching engine from rules and weights to semantic matching, embeddings, and vector search. Build rigorous precision/recall evaluation, entity resolution across 8M+ NPI records, and a reliable lightweight data platform using healthcare data. Independently own ingestion, enrichment, modeling, serving, and architecture decisions while improving recommendation quality and supporting customer-facing reporting.

Own and improve Tessellate’s physician-to-organization matching engine by evolving rules-based matching toward semantic matching, embeddings, and vector search. Establish precision, recall, and evaluation practices; resolve healthcare organization identities across millions of records; and maintain a reliable, lightweight data platform. The role requires strong Python and SQL, applied machine learning expertise, production recommendation or ranking experience, and end-to-end ownership in an early-stage startup.

Build and ship physician-facing product features across a mobile-first React frontend and Python backend. Extend APIs, evolve schemas, and deliver tested production work while collaborating on architecture and engineering standards on a small team.