Text2SQL

Schema mapping system using Word2Vec for semantic column naming and LLMs to generate SQL views with joins, type casting, and aggregations — enabling natural language access to biomedical databases. Achieved 82% precision and 73% recall.