ShopSmart AI Shopping Assistant
Intro
ShopSmart was developed during the Personalized Agents Hackathon in NYC, hosted by Lightning AI and Validia, where the focus was on building autonomous agents that are both effective and secure. I designed ShopSmart as a decision-support agent for online shopping. Instead of overwhelming users with options, it narrows down choices and highlights the safest, best-value products using real-time data and a custom scoring system. The system integrates: - Python + Gradio to build an interactive, user-facing interface - SerpApi (Google Shopping) for live, cross-retailer product data - A custom ranking algorithm that balances price, ratings, and review volume - A risk detection layer to flag low-quality or potentially misleading listings - Lightning AI + Clawdbot to structure agent workflows and support scalable decision logic This project explores how AI agents can move beyond simple search and into real-world decision-making support, where accuracy, trust, and risk awareness matter.
