DEFENCE RESEARCH UDHAYAKUMAR R

An interactive single-page web application (SPA) designed to visualize and analyze information on modern hybrid warfare tactics and India's strategic countermeasures. Developed as submission for Combat Vehicles Research & Development Establishment (CVRDE), Avadi internship.

• Interactive visualization of hybrid warfare tactics
• Strategic countermeasures analysis for Indian defense
• Single-page application with modern web technologies
• Submitted to CVRDE for internship evaluation

Co-developed an AI system with Rohit Arvind S to identify medicinal plants and enhance supply chain transparency

• Integrated CNNs and IoT sensors for real-time environmental monitoring
• Achieved 85% accuracy in plant identification
• Reduced overharvesting and ensured traceability via geolocation mapping
• Published in IEEE conference proceedings

Co-developed an AI system with Rohit Arvind S to identify medicinal plants and enhance supply chain transparency.

• 85% accuracy in plant identification using CNNs
• 35% improvement in classification accuracy
• React.js interface for user-friendly interaction
• 20% reduction in misidentification errors
• Supports sustainable herbalist practices

Created an image-based authentication system integrating machine learning for secure user verification

• Used Python and OpenCV for image processing
• Focus on facial recognition and data encryption
• Designed scalable solution for secure access control systems
• Open source implementation available on GitHub

Developed an AI-powered conversational web app to democratize access to Argo oceanographic data through natural language queries.

• Built immersive ocean-themed UI with React-style frontend concepts and Tailwind CSS.
• Integrated Framer Motion-inspired animations and smooth page transitions.
• Implemented interactive visualizations with Chart.js and Leaflet.js for ocean maps and float tracking.
• Added dynamic customizable light/dark theme system with seamless transitions.
• Simulated backend API and NLP for prototyping AI conversational user experience.
• Co-Contributor: Prasana K

A modern web app using satellite Sentinel-5P/TROPOMI + ERA5 data for real-time ground-level ozone (O₃) and nitrogen dioxide (NO₂) short-term forecasting with AI models.

• Interactive Leaflet map visualizations of air pollution spatial forecasts.
• ConvLSTM and Transformer based machine learning model selections.
• Dark/light mode with smooth animations and air quality thematic design.
• AI chatbot integrated for natural language air quality queries.
• Full React + Tailwind CSS frontend with FastAPI simulated backend.