Next-Generation Web Apps with AI

I’ve built several web-based applications that integrate AI to explore different aspects of interactive experiences, search, and communication. Each project was an experiment in combining AI capabilities with user-friendly interfaces to demonstrate how machine learning models can enhance engagement and functionality in web applications.

Rill Search is a lightweight AI-powered search engine that blends traditional indexing with modern language models. Built with Flask, Whoosh, and OpenAI’s ChatGPT, it includes web crawling, indexed search, spelling corrections, and AI-generated summaries for search results. The combination of LLM-generated content previews and smart crawling techniques ensures that search results are not only more relevant but also more informative.

Chat and React is a distributed chat system featuring a hub-and-spoke architecture. It allows multiple specialised chat channels to connect to a central registry, with a front-end built in React and TypeScript. The system includes a Privacy Advisory Channel, where an AI-powered bot provides contextual privacy advice based on user conversations. The modular architecture supports extensibility, real-time messaging, and a modern UI with markdown formatting and theme customization.

Who’s That Pokémon? is an AI-powered guessing game developed with Streamlit and OpenAI’s GPT models. Players receive AI-generated hints to identify a randomly chosen Pokémon, while a judge model evaluates their accuracy. The game features interactive chat, a scoring system, session persistence, and detailed gameplay statistics.