Research and industry roles across computer vision, deep learning, and federated learning.
Computer Vision and Pattern Recognition Lab, IIT Ropar
Junior Research Fellow - Federated Learning and Deep Learning
Worked on federated learning for aerial scene classification under the supervision of Dr. Santosh Kumar Vipparthi.
Built FedRWA for 12 clients and 313 disjoint classes, improving accuracy by 18.6% and reducing client variance by 57.8% across three backbones.
Image Processing, Security & Analytics Labs, IIT Ropar
Research Intern - Applied Deep Learning
Worked on applied computer vision problems spanning camera forensics, disaster imagery, and human-action understanding.
Developed SPAIR-Swin across four camera datasets, improved disaster-image F1 scores by 2.41 to 9.93% with ConvSFNet, and advanced yoga-pose recognition with HYPC-Net.
Machine Learning Intern - 3D Computer Vision
Worked on indoor point-cloud understanding with PointNet++, improving classification performance by 25%.
Improved IoU for underrepresented classes by 30 to 45% and automated data and evaluation workflows to reduce processing time by 30%.