Built an MCP-based Text-to-SQL + RAG pipeline that retrieves Indian Ocean ARGO float data by integrating a MySQL database with ChromaDB vector search, enabling context-aware query results. Designed the backend workflow: data preprocessing, embedding generation, hybrid retrieval, and MCP orchestration to support oceanographic research use-cases.