ShopShield AI is an AI-powered product recommendation system that helps users discover relevant products through natural language queries. The project uses Retrieval-Augmented Generation (RAG) with vector search to retrieve product information and generate personalized recommendations.
- User authentication (username & password)
- AI-powered product recommendations using Google Gemini
- Retrieval-Augmented Generation (RAG) pipeline
- Product retrieval using vector embeddings
- Semantic product search using ChromaDB
- REST APIs built with FastAPI
- Modular backend architecture for future feature expansion
- Python
- FastAPI
- Google Gemini API
- LangChain
- HuggingFace Embeddings
- ChromaDB
- PostgreSQL
- Product information is converted into vector embeddings.
- Embeddings are stored in ChromaDB.
- User submits a query through the API.
- Relevant products are retrieved using semantic search.
- Retrieved context is passed to Gemini.
- Gemini generates personalized recommendations.
The recommendation workflow consists of:
- Chunking product information
- Generating embeddings
- Vector storage in ChromaDB
- Similarity-based retrieval
- Context generation using Gemini
POST /recommendReturns personalized product recommendations.
POST /searchPerforms semantic product search.
git clone https://github.com/tailormst/Ecommerce-Amazon-Recommendation-Assistant.gitpip install -r requirements.txtCreate a .env file:
GOOGLE_API_KEY=your_api_keyuvicorn app.main:app --reloadThrough this project, I gained practical experience with:
- FastAPI backend development
- REST API design
- Retrieval-Augmented Generation (RAG)
- Embeddings and vector databases
- ChromaDB integration
- Semantic search systems
- Gemini API integration
- Chat history support
- Product comparison features
- Recommendation analytics
- Multi-modal product search
Mohammed Saifuddin Tailor
GitHub: https://github.com/tailormst