A collection of innovative solutions I've built, from AI-powered applications to creative mobile experiences.
Food ML Analysis App that allows users to scan packaged or open food items to instantly analyze health risks and receive personalized alternative recommendations. The app integrates machine learning with Retrieval-Augmented Generation (RAG) to interpret ingredient data, leverages the Gemini API for intelligent analysis, and uses Firebase for secure authentication and data management. Used by 100+ students.

Website that automatically indexes GitHub repositories using vector embeddings and provides intelligent, context-aware development guidance through a LLM pipeline. Features real-time chat interface, webhook-based auto-indexing, and microservices architecture deployed across Vercel and Render. Also provides bots that test app just like users do in beta version to gain feedback.

Desktop App for analyzing user behavior and understanding repetitive tasks by LLM to automate them. Multi-threaded OCR and Whisper for each FPS being recorded. Used smollm2:1.7b as offline LLM, data streaming of screen between Flutter and Python subprocess. Designed modular for cross-platform Windows production-ready, macOS/Linux compatible.
Pre-order app for food ordering before reaching the canteen. Features real-time updates on meals, payment gateway integration, and time-tracking functionality.
Movies updates and trailers viewing app providing information about actors, movies and their trailers similar to IMDB functionality.
Time and place capture app that can capture images of moments and store them in memory with location data for easy access and retrieval.
Cloud-based note-taking application for keeping all notes in the cloud with access from anywhere. Features synchronization across devices.