Built a full-stack platform that turns any public website into an AI-queryable knowledge base. Using FastAPI and Python, the system performs multi-page web crawling with configurable depth and converts unstructured content into structured JSON and Markdown. The data is indexed in a vector database and combined with LLMs via Ollama to enable context-aware question answering using retrieval-augmented generation, ensuring responses stay grounded in source content. A React.js frontend provides a real-time interface for controlling crawls and querying results, effectively allowing users to interact with any website like a domain-specific AI assistant.

Developed a flexible point calculation system that allows users to define custom game rules, add participants, and automatically compute scores in real time. The application determines winners and losers based on dynamic scoring logic, making it adaptable to various game formats. Designed with a user-friendly interface, the system simplifies score tracking and eliminates manual calculations, providing a seamless experience for managing competitive gameplay.

Built a bill-splitting application that simplifies group expense management by allowing users to create groups, track frequent venues, and itemize orders with individual pricing. The system intelligently splits the total bill based on what each person consumed, ensuring accurate and fair cost distribution. Designed for real-world usability, it eliminates manual calculations and confusion in shared expenses, providing a seamless way to manage group payments.