This project is an AI-powered multi-tenant perfume recommendation system that uses LLMs (GPT-4o) and semantic vector search to provide personalized fragrance discovery. It features a Django-based backend with isolated shop data and an interactive chat interface with session-based memory for customers.  Core Components • AI Engine: Uses a hybrid logic combining similarity search with local inventory through an embedding pipeline.  • Tech Stack: Built with React and Tailwind CSS on the frontend, and Django REST Framework with PostgreSQL (pgvector) on the backend.  • Dashboards: Includes a Shop Admin for product management and a Super Admin for tenant management and system-wide analytics.  • Infrastructure: Fully containerized with Docker and deployed on Azure Cloud using automated CI/CD pipelines.