About Me
I am Abu Mukaddim Rahi, a researcher and developer specializing in deep learning, medical imaging, and natural language processing. Currently, I work as a Research Assistant at North South University, where my work focuses on semantic embeddings, continual learning, and medical image analysis. My research has resulted in multiple publications in top-tier conferences and journals, including Frontiers in Medicine, Results in Engineering, and international AI conferences.
Alongside my research, I have been with Hash Solution for nearly two years, starting as a Front-End Engineer and later contributing to both front-end and backend development as a Junior Developer to support AI integrations. I design and deploy scalable front-end and full-stack applications using React.js, TypeScript, Tailwind CSS, and Go Language, while integrating AI-driven solutions into production systems.
Experience
Research Assistant
CyberX Mentor, Dhaka
Conducted research in deep learning, NLP, and medical imaging, contributing to model optimization, dataset creation, and publications in Q1 journals and IEEE/Scopus conferences while supporting AI-driven projects through benchmarking and experimentation.
Junior Full Stack Developer (Remote)
Hash Solution, Dubai
Leading front-end and full-stack development (Remote) using React, JavaScript, TypeScript, Tailwind CSS, and Go, integrating AI-driven solutions and mentoring interns to ensure scalable, high-performance web systems.
Research Assistant
North South University, ECE Department
Working under the supervision of Md. Shahriar Hussain on medical imaging, deep learning, NLP, and machine learning, with a focused emphasis on advanced research
Publications & Preprints
Not all of my conference papers are currently listed on Google Scholar, as most of my publications are from 2025 and may take some time to appear. To view my papers, please scroll down and click the links provided. Google Scholar.
2024
"Custom Lightweight CNN and Data-Efficient Models for Efficient and Fast Convergence in Lung Cancer Classification,"
In The 27th International Conference on Computer and Information Technology (ICCIT 2024). (IEEE Xplore, Scopus)
2025
"Multi-Model Deep Learning Approach for Multi-Class Brain Tumor Detection from MRI,"
In The 33rd International Conference on Software, Telecommunications, and Computer Networks (SoftCOM 2025). (IEEE Xplore, Scopus)
"A Hybrid Deep Learning Model for Accurate Chest X-Ray Disease Classification,"
In The International Conference on Computing and Machine Intelligence (ICMI 2025). (IEEE Xplore, Scopus)
"Benchmarking Large Language Models for Sarcasm Detection in Hindi Romanized Texts"
In The International Conference on Electrical, Computer and Energy Technologies (ICECET 2025). (IEEE Xplore, Scopus)
"Fine-Tuning LLaMA-3.2-3BIT on Bengali Hate Speech Dataset for Sentiment Classification,"
In Conference on Secure IoT, Assured and Trusted Computing (SaTC 2025). (IEEE Xplore, Scopus)
"Automated Stroke Prediction Using Machine Learning Approach,"
In International Conference on Electrical, Computer and Energy Technologies (ICECET 2025). (IEEE Xplore, Scopus)
"Performance optimization of Predictive Analytics Algorithm for Accurate Cyber Threat Forecasting"
In International Conference on Electrical, Computer and Energy Technologies (ICECET 2025). (IEEE Xplore, Scopus)
Journal Paper
"Multimodal Encoder-Decoder Framework for Anatomically Precise Fracture Caption Generation from Medical Imaging"
In Frontiers in Medicine, Volume 12, 2025 . top journals worldwide in medicine (Q1 Ranked, Scopus & WoS)
"CerevianNet: Parameter-Efficient Multiclass Brain Tumor Classification Using Custom Lightweight CNN"
In Frontiers in Medicine, Volume 12, 2025 . top journals worldwide in medicine (Q1 Ranked, Scopus & WoS)
"Robust Sentiment Classification of Bangla Social Media Text Using Fine-Tuned LLaMA Model and Voting-Based Inference"
In Machine Learning: Science & Technology, IOP Science . (Q1 Ranked, Citescore:7.7)
"MultiModal Meme Hateful Detection: A Comparative Study on Visual Language Models and Fusion Architectures 2025"
In Results in Engineering (Elsevier) . (Q1 Ranked, Impact Factor:7.9, Scopus & WoS)
Projects
E-Invoice Solutions
• Contributed to the development of E-Invoice Solutions, a cloud-based platform for
managing digital invoices,
customers, and organizations. Implemented features such as invoice processing, customer
management, and API
integration to streamline financial operations and compliance.
• Technology: Python, Go Programming Language, TypeScript, JavaScript.
Find Me a Garage
• Contributed to the development of E-Invoice Solutions, a cloud-based platform for
managing digital invoices,
customers, and organizations. Implemented features such as invoice processing, customer
management, and API
integration to streamline financial operations and compliance.
• Technology: Python, Go Programming Language, TypeScript, JavaScript.
Disaster Rescue AI
• Disaster Rescue AI: Advanced AI for Finding Alive People in Disasters and AlertSystem,
developed
an AI-powered system for real-time person detection in disaster scenarios using YOLOv8
and YOLOv10,
optimized for speed and accuracy. Supervised by Dr. Mohammad Ashrafuzzaman Khan,
Associate Professor,
ECE Dept, North South University, as my senior year project. North South
University.
• Technology: Python, Django, Tensorflow, Google Colab, Kaggle.
• Senior Design Project, CSE499.
Lips Movement Detection
• Developed a real-time computer vision application for detecting and tracking lip
movements,
supporting speech recognition, lip reading, and facial expression analysis. Trained deep
learning
models, built the detection pipeline, and evaluated performance across diverse
scenarios, providing
a practical tool for researchers and developers.
• Technology: Python, Django, Tensorflow, Google Colab, Kaggle.
• Junior Design Project, CSE299.
Achievements
Explore my certificates, recognitions, and key milestones. View All Achievements
Machine Learning Specialization
The Machine Learning Specialization on Coursera by Stanford Online and Andrew Ng, successfully finishing multiple quizzes and assignments, and earned the official certificate.
High-Quality Reviews
I received a Certificate of Appreciation from ICECET 2025 for High-Quality Reviews, IEEE International Conference, for my valuable contribution as a reviewer.