QueryPilot AI | Retail Analytics Copilot | github.com/fmahtab/querypilot-ai
Built a business analytics copilot that answers natural-language questions using business documentation. Implemented a RAG workflow that retrieves relevant context and generates grounded answers with source attribution. Added short-term session memory for conversational continuity and durable memory for retaining relevant context across sessions. Developed the FastAPI service, /ask request and response schemas, PostgreSQL data layer, SQLAlchemy models, and Alembic migrations. Built an evaluation suite covering request routing, retrieval quality, groundedness, source attribution, questions requiring database access, and unsupported-query handling.


.jpg)
