An AI-powered VS Code extension that turns source code into conversational audio content. It extracts relevant code context, generates dynamic podcast-style scripts using Gemini, and synthesizes developer-friendly audio summaries directly inside VS Code.

A multilingual AI healthcare assistant designed to provide preliminary health information and localized medical education. Built with a React frontend and FastAPI backend using Gemini and Retrieval-Augmented Generation for contextual responses.

An open-source WhatsApp integration bridge for Google Agent Development Kit that routes real-time user messages to autonomous AI agents using event-driven webhook handlers, session state management, message serialization, and automated response callbacks.

A browser-based AI-focused video editor built for modern creators. Features include media importing, transforms, masking, blending modes, captions, and browser-based video export powered by FFmpeg WebAssembly.

A document question-answering application that combines PDF ingestion, vector search, retrieval-augmented generation, and LLMs to let users interact with their documents conversationally.

An experimental environment for building, testing, and observing AI browser agents. Uses browser automation, remote browser sessions, VNC-based visualization, and agent tooling to explore practical autonomous browser workflows.

An experiment exploring whether small local language models can make real-time game decisions. Captures the game screen, detects obstacles using computer vision, and evaluates actions through locally running models such as Tev1 and Ollama.