Portrait

I am Md Mostafijur Rahman, a final year Ph.D. Candidate at The University of Texas at Austin advised by Prof. Radu Marculescu. You can view my CV here and reach me at mostafijur.rahman [at] utexas.edu.

I design and build AI systems that make medical imaging and computer vision more efficient, trustworthy, and accessible in low-resource settings. My research treats practicality as a first-class concern, focusing on computational efficiency, data efficiency, robustness, and generalization through efficient architectures, foundational models, robust training, generative, and multi-modal learning methods for 2D/3D segmentation, classification, synthesis, translation, and denoising. I've developed methods such as: LoMiX (NeurIPS ’25), a plug-and-play supervision module for stable multi-scale learning; EfficientMedNeXt (MICCAI ’25 Highlight), a 3D medical image segmentation architecture that delivers end-to-end efficiency; EffiDec3D (CVPR ’25 Highlight), an optimized 3D decoder that boosts accuracy at lower cost; EMCAD (CVPR ’24), an efficient multi-scale convolutional-attention decoder that strengthens 2D segmentation; and CASCADE (WACV ’23), an early cascaded attention-based decoder for medical image segmentation.

During my Ph.D., I worked as a machine learning research intern at GE HealthCare, the National Institutes of Health (NIH), and Bosch AI. In the past, I worked as a Senior Software Engineer at Samsung Research Bangladesh and spent four years as a faculty (Assistant Professor and Lecturer) at the University of Barishal, Bangladesh.

📣📣📣 I am currently on the job market, looking for tenure-track faculty or research roles.

News

Selected Publication [Full List]

Awards

Teaching Experience

Professional Experience

Advising & Mentoring

Academic Services

Presentations

Contact

Address

EER 5.838B
2501 Speedway
Austin, TX
United States
78712

Email

Academic
mostafijur.rahman [at] utexas.edu

Other Use
mostafij.csebu [at] gmail.com

Copyright © 2024 Md Mostafijur Rahman. All Rights Reserved.