PROJECTS & PUBLICATIONS
- EmojiCamera: Facial Expression Detection AI
Led the development of a real-time AI system that detects facial expressions and maps them to corresponding emojis. By leveraging MobileNetV3 with attention mechanisms, the model achieved a 75% accuracy. Additionally, I co-authored a research paper, “Lightweight Facial Expression Recognition Models for Low-Cost Computing Environments”, which was presented at the 2025 IEEE Conference on Artificial Intelligence.
- Flight Subscription Service
Developed a flight deal subscription platform that integrates REST APIs for real-time flight data retrieval. The system allows users to sign up for alerts and search for deals, improving user engagement. Built with React (frontend), Node.js (backend), and MySQL (database), the platform optimizes performance and scalability.
- Advanced Car Bidding System
Designed a real-time car auction platform with a Django backend, React frontend, and MySQL database. The system features secure user authentication, dynamic bidding interfaces, and Docker-based deployment on GCP. My contributions improved the platform’s usability, security, and scalability.
- Vaccine Companies Stock Price Forecast During COVID-19
Developed a time series forecasting model using ARIMA and SARIMA to predict stock price trends for Pfizer, Johnson & Johnson, and Moderna during the COVID-19 pandemic. By collecting official CDC data (daily infections, deaths, vaccination rates, etc.) and applying differencing techniques, the model was trained to capture stock price movements. However, due to the high volatility of stock markets and numerous external factors, the prediction accuracy remained around 50%, highlighting the complexity of relying solely on pandemic-related data for financial forecasting.