Ayon Adhikary
Bachelor of Science in Computer Science & Engineering
East West University
Project
Developing a GAT-based software defect prediction system using NASA/PROMISE datasets (JM1, KC2, and PC1), with automated C/C++ metric extraction, defect probability prediction, SHAP-based explainability, and automated CSV dataset generation for future mod
Project
Developed and evaluated YOLO-based object detection models for detecting pedestrians, vehicles, and animals in thermal imagery using the FLIR dataset. Compared YOLOv10s, YOLOv12s, and YOLOv26s using precision, recall, F1-score, and mAP metrics.
Project
Developed aerial vehicle detection models using supervised and self-supervised learning on oriented bounding box datasets. Explored STAC, BYOL, and SimCLR for label-efficient object detection under limited labeled-data settings.
Project
Developed a full-stack real-time auction platform featuring secure authentication, live bidding, bid history, wishlists, automated auction scheduling, and an administrative analytics dashboard. Integrated Socket.io for real-time communication and Gemini A