Meheraj Hasan
Machine Learning Researcher · AI Engineer
Deep learning for medical image analysis — classification, detection, and segmentation across ultrasound, fundus, histopathology, and dermoscopic imaging, with a focus on real-world deployment under limited annotation and real-time constraints.
A researcher focused on deep learning for medical image analysis, spanning classification, detection, and segmentation across ultrasound, fundus, histopathology, and dermoscopic imaging. Studying architectures that remain robust under real-world constraints — limited annotation, class imbalance, multi-vendor imaging hardware, and real-time inference requirements. Particularly interested in extending this work toward federated and privacy-preserving learning, enabling training on distributed clinical data without centralization.
Currently an AI Engineering Intern at Samsung R&D Institute Bangladesh (SRBD), developing real-time detection and segmentation models for ultrasound-based diagnostic pipelines under strict latency and FPS constraints. Also co-authoring research on a role-based crowdsourcing framework for Bengali lexicon development, in collaboration with researchers at the University of Trento and DataScientia, Italy.
Previously a Research Intern at the Advanced Machine Intelligence Research Lab (AMIR Lab), advised by Prof. Dr. Muhammad Firoz Mridha, working on medical image analysis and applied deep learning.
Medical Image Analysis Research
Co-authored publications on medical image analysis spanning breast, retinal, brain, skin, and blood-smear imaging — accepted, under review, or ongoing at international conferences and journals.
Real-Time Ultrasound at Samsung R&D
Developing real-time detection and segmentation models for thyroid ultrasound diagnostics at Samsung R&D Institute Bangladesh (SRBD), balancing clinical accuracy against FPS and latency constraints.
Bengali Lexicon & Universal Knowledge Core
Co-authoring a role-based crowdsourcing framework for Bengali lexicon development, compatible with the Universal Knowledge Core (UKC), in collaboration with the University of Trento and DataScientia, Italy.
Honorable Dean's Award (5×)
Recipient of the Honorable Dean's Award for five consecutive semesters, Faculty of Science and Technology, American International University–Bangladesh.
Also exploring:Cross-device/cross-vendor generalization in medical imaging · Bioinformatics and computational biology applications of deep learning · Multimodal learning for clinical decision support · Digital pathology and computational histology · LLMs and retrieval-augmented systems for clinical reasoning (emerging interest)
B.Sc. in Computer Science and Engineering
Sept 2022 – Aug 2026American International University-Bangladesh (AIUB)
Dhaka-1229, Bangladesh
Recipient of the Honorable Dean's Award for five consecutive semesters and full Merit-Based Academic Scholarship (Faculty of Science and Technology).
Intern Engineer (AI)
March 2026 – PresentSamsung R&D Institute Bangladesh Ltd. (SRBD)
AI Solution Team, Medical Solution Group · Dhaka-1205, Bangladesh
Supervisor: Muhammad Rakeeb, Assistant Principal Engineer, SRBD
- Build and analyze medical AI project pipelines using ultrasound imaging.
- Conducted R&D on SOTA models and architectures.
- Proposed novel architectures in terms of accuracy and efficiency.
- Write scalable and maintainable model training and integration scripts.
Research Intern
Jan 2026 – May 2026Advanced Machine Intelligence Research Lab (AMIR Lab)
Medical Image Analysis & Applied Deep Learning · Dhaka, Bangladesh
Supervisor: Prof. Dr. Muhammad Firoz Mridha
- Conducted research on advanced machine intelligence and deep learning methodologies.
- Worked on computer vision and large language model architectures for real-world research problems.
- Collaborated with researchers on experimental design, model evaluation, and result analysis.
International Journals
Meheraj Hasan, Fariha Alam Urbana, Mahamodul Hasan Mahadi, Firoz Ahmed, Md. Shabiul Islam, Sadia Sharmin. “PoxCSAF-Net: Cross-Scale Attention Fusion for Privacy-Aware Federated Classification of Mpox and Visually Similar Skin Lesions.” Frontiers in Digital Health, 2026.
Meheraj Hasan, Sultanul Arifeen Hamim, Muhammad Firoz Mridha. “RetiCut-Net: Source-Specific Ordinal Cut-Points for Multi-Site Diabetic Retinopathy Grading.” International Journal, 2026.
Saiful Islam Oni, MD Sazzad Khan, Meheraj Hasan, Shams Pahlowan Soad, Khandaker Tabin Hasan, Hadi Mahmoud Yousef Khalilia, Fausto Giunchiglia. “Ben-LKC: A Language-Independent Role-Based Crowdsourcing Framework for UKC-Compatible Bengali Lexicon Development.” International Journal, 2026.
International Conferences
Meheraj Hasan, Nazifa Tabushum Turaba, Ainea Esrat Esika, Sanjid Raaz, Sultanul Arifeen Hamim, Muhammad Firoz Mridha. “HemaVision-Net: A YOLOv11-Based Framework for Severely Occluded Blood Smears.” International Conference on Data Analytics for Business and Industry (ICDABI), 2026.
Sumshun Nahar Eity, Meheraj Hasan, Hasan Mahmud, S. M. Mehedi Hasan, Tanisha Fairooz, Kazi Fattah Afnan, Rifath Mahmud. “A Comparative Performance Analysis of State-of-the-Art Deep Learning Architectures for Malting Barley Disease Classification.” International Conference on Electrical and Computer Engineering (ICECE), 2026.
Meheraj Hasan, Taharat Muhammad Jabir, Md. Rehab Ansary Yasin, Mohammed Samir, Sharfuddin Mahmood. “DINO-MIL: An ROI-Based Feature Embedding Framework for Multiple Instance Learning in Breast Mass Classification.” International Conference on Quantum Photonics, Artificial Intelligence & Networking (QPAIN), pp. 1–6, 2026.
Saiful Islam Oni, Meheraj Hasan, Arpon Paul Amit, Shahnaj Parvin, Farzana Bente Alam, Kamruddin Nur. “Detection and Classification of Diabetic Retinopathy Using EfficientNetB0: A Lightweight Model with Transfer Learning Approach.” International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), pp. 1–6, 2026.
AI-Assisted Diagnostic System for Medical Ultrasound Imaging
Affiliated with AI Solution Team, Medical Solution Group, Samsung R&D Institute Bangladesh Ltd. (SRBD)
March 2026 – Present
- Built and cleaned segmentation and detection datasets from multi-vendor ultrasound machine images.
- Studied, fine-tuned, and added novelty to detection models to meet real-time FPS and performance benchmarks.
- Developed segmentation models with novel improvements to meet clinical accuracy requirements on freeze-frame ultrasound images.
- Contributed to an automated structured reporting pipeline (TI-RADS style) generated from segmentation outputs.
Expert-in-the-Loop Medical Image Annotation Tool
Affiliated with AI Solution Team, Medical Solution Group, Samsung R&D Institute Bangladesh Ltd. (SRBD)
October 2026 – November 2026
- Developed a desktop tool for clinical experts to annotate medical images and video frames for project-specific datasets.
- Implemented OpenCV-based bounding-box propagation across sequential frames with confidence scores, where a correction on any frame auto-updates the rest.
- Built a grid view comparing original images, expert annotations, and predictions from user-selected local ONNX models.
- Designed a role-based review workflow: annotators approve or delete auto-generated boxes individually or in batches, doctors approve, reject, or correct annotators' labels, and full undo/redo history is supported.
- Added image enhancement controls (brightness, contrast, hue, saturation, CLAHE, and single RGB channel view) to improve nodule visibility.
Drinking Water Safety Prediction using Machine Learning
End-to-End ML System Development
Sep 2025 – Nov 2025
- Developed an end-to-end machine learning system to predict drinking water safety using physicochemical features.
- Deployed the trained model as a public web application using Gradio and Hugging Face Spaces.
TravelEaseV2
Unified Ticket Booking System
- Designed and developed a backend-focused ticket booking system using ASP.NET Core Web API.
- Implemented RESTful APIs following clean 3-tier architecture (API, BLL, DAL).
- Built complete booking workflow including seat reservation, availability checking, and booking status management.
- Implemented payment processing with fare calculation, discount rules, and payment validation.
- Applied repository pattern, service layer abstraction, and AutoMapper for clean code separation.
- Implemented JWT-based authentication and role-based authorization.
- Built a full-stack web application for tracking income, expenses, debts, savings, and financial goals.
- Implemented user authentication and state management using sessions and cookies.
- Developed asynchronous data handling using AJAX and JSON for seamless user interaction.
- Generated real-time interactive financial charts, analytics dashboards, and automated reports.
- Developed a centralized ticket booking system for buses, trains, airplanes, and launches.
- Supports user registration, ticket purchasing, vehicle owner management, and customer support workflows.
Research Member
AIUB Research and Development Club
Actively participated in technical workshops and collaborative research projects focused on modern software development technologies, data-driven systems, and applied computing methodologies.
Also proficient in software systems & engineering:ASP.NET Core / MVC, Windows Forms, PHP, Flutter, SQL Server, MySQL, HTML/CSS/JavaScript
References available upon request.
Prof. Dr. Muhammad Firoz Mridha
Professor & Department Head (Undergraduate Program)
Department of Computer Science and Engineering
American International University-Bangladesh (AIUB)
Prof. Dr. Khandaker Tabin Hasan
Professor & Associate Dean
Faculty of Science & Technology
American International University-Bangladesh (AIUB)
Muhammad Rakeeb
Assistant Principal Engineer
Medical Solution Group
Samsung R&D Institute Bangladesh Ltd. (SRBD)
Full referee direct contact details are provided on the formal Curriculum Vitae PDF download.
Open to discussions regarding PhD opportunities, collaborative deep learning research in medical imaging, or engineering consultations.