Developed real-time disaster monitoring platform processing Bluesky tweets through ML classifier and Ollama LLM integration for automated event detection and categorization.

• Designed workflow pipeline for preprocessing tweet data and classfying using XGBoost classifier
• Integrated local Ollama LLM with custom prompting for disaster metadata extraction including event type, geographic location, and urgency level
• Built React UI dashboard displaying real-time disaster alerts with interactive filtering and maps

Created a version-controlled note-taking platform with real-time synchronization for seamless markdown document management.

• Developed CLI tool using Commander.js for efficient note synchronization and offline workflows
• Integrated LangChain AI for document summarization and QA chatbot interactions with user content
• Built scalable PostgreSQL database with Supabase for robust data persistence

Developed full-stack expense tracking application with receipt image upload and OCR data extraction.

• Implemented Tesseract.js OCR engine to automatically extract expense data from receipt images
• Built RESTful API with Express.js and Multer for file uploads and expense management
• Created responsive interface with expense history, sorting, filtering, and receipt display features

Developed interactive educational platform teaching STEM concepts to younger students through engaging experiences.

• Built responsive frontend with Three.js integration for dynamic 3D visualizations optimized for performance
• Integrated OpenAI API to generate real-time educational content and adaptive quiz questions