Selected Work

Developed a smart building safety system that detects fire outbreaks and dynamically routes occupants to the safest exit. The system features an admin dashboard with a live floor map visualizing real-time sensor data and calculated escape routes. Engineered the dynamic routing algorithm that processes current sensor values to calculate the safest path, paired with physical LED strips in corridors to guide evacuees.

Developed a microservices platform connecting food donors with NGOs to streamline surplus food listing and pickup. Built Go backend microservices for authentication, user profiles, food listings, and claims using PostgreSQL and JWT tokens, containerizing each service with Docker. Added background checks and database locking to automatically expire old listings and prevent double-claiming.

Developed a deepfake detection system utilizing Convolutional Neural Networks (CNNs) to accurately classify whether facial images are real or AI-generated. Gained deep practical experience in CNN architecture, feature extraction, and model training to identify subtle artifacting left by generative models.

Engineered a smart parking solution for commercial buildings and urban areas, allowing users to book and manage parking spots via a mobile application. The system automates the entire parking lifecycle, ensuring efficient space utilization and maintenance. The research paper detailing this IoT-driven system was published in Springer as part of the ICT4SD 2025 international conference.