🛒 Amazon AI Shopping Assistant

An AI-powered shopping assistant that recommends Amazon products using Machine Learning and Large Language Models.

Built with FastAPI, Streamlit, and Hugging Face FLAN-T5, the application combines customer review analytics with Generative AI to recommend the best product based on thousands of Amazon reviews.

Developed as the deployment phase of the Ironhack AI Engineering Bootcamp.

---

🚀 Live Demo

Application:
https://ai-shopping-assistant-ih.streamlit.app

API Documentation:
https://ai-shopping-assistant-production-14df.up.railway.app/docs

---

How It Works

User Question
      ↓
Intent Detection
      ↓
Retrieve Best Product
      ↓
Retrieve Customer Evidence
      ↓
FLAN-T5
      ↓
Shopping Recommendation

---

Technologies

• Python
• FastAPI
• Streamlit
• Hugging Face Transformers
• FLAN-T5 Small
• DistilBERT
• TF-IDF + K-Means
• Railway
• Streamlit Community Cloud

---

Features

• Natural language shopping questions
• AI-generated product recommendations
• Customer review analysis
• FastAPI REST API
• Streamlit frontend
• Fully deployed cloud application

---

Repositories

Deployment Repository
https://github.com/Martigol2/AI-Shopping-Assistant

Original NLP Project
https://github.com/Martigol2/IRONHACK-AI-PROJECT3

---

Team

• Felipe Martignon
• Casilda Gil de Santivañes

GitHub:

• https://github.com/Martigol2
• https://github.com/Casildagsf