projects | Mohammad Junayed Hasan

Featured Research

OptimCLM

IJMI 2025

Optimizing clinical language models via knowledge distillation, pruning, and quantization. Achieved 22.9× compression with 98% performance retention.

Code Paper

HadaSmileNet

IEEE ICDM 2025

Hadamard fusion of handcrafted and deep-learning features for genuine smile recognition. Oral + Poster at ICDM 2025.

Code Paper

Bridging Quantum-Classical ML

IEEE TQE (Under Review)

First knowledge distillation framework bridging classical neural networks with quantum circuits. 12 preprint citations.

Code Preprint

Shadow Loss

CVPR 2026 (Under Review)

Memory-linear deep metric learning reducing complexity from O(N²) to O(N) while accelerating convergence 1.5-2×.

Preprint

industry

Blood Horizon Predictor

Predictive blood utilization management system for enhanced transfusion medicine operations at Mayo Clinic

HER2 Domain Adaptation

Extending HER2 biomarker detection from breast cancer to 7 cancer types using domain adaptation and transfer learning

research

Bridging Quantum-Classical ML

Knowledge transfer from classical to quantum neural networks via knowledge distillation (IEEE TQE - Under Review, 12 preprint citations)

DeepMarkerNet

Multi-task transformer framework for spontaneous smile recognition using Duchenne Marker supervision (PRL 2024)

HadaSmileNet

Hadamard fusion of handcrafted and deep-learning features for genuine smile recognition (IEEE ICDM 2025)

OptimCLM

Optimizing clinical language models via knowledge distillation, pruning, and quantization (IJMI 2025)

ParaCycle

Reinforcement learning framework for reference-free machine translation using bidirectional paraphrase consensus

Shadow Loss

Memory-linear deep metric learning reducing complexity from O(N²) to O(N) (CVPR 2026 - Under Review)

NCD Detection Framework

Novel framework for noncommunicable disease detection via prompt engineering and domain knowledge integration (AEJ 2025)

Plant Leaf Disease Detection

Deployable deep learning for cross-domain plant disease detection via ensemble learning, knowledge distillation, and quantization

Occupational Stress Detection

Early detection of occupational stress using machine learning and large language models with ~90% accuracy

Life Satisfaction Prediction

Predicting life satisfaction using machine learning, explainable AI, and novel tabular-to-text NLP algorithms

github

Disability Care

Web application connecting people with disabilities to healthcare services through an accessible interface

LifeEmbedding

Cloud-native system to model human life trajectories as embeddings with retrieval-augmentation

Valentines Blossoming Flower

Interactive single-page web application with animated UI elements and audio-visual effects

Quantum Machine Learning

Comprehensive tutorials and hybrid quantum-classical models using Qiskit and PyTorch

LLendify

AI-powered loan eligibility system using LLMs to analyze bank statements and provide instant recommendations

CrazyCrawler

Web search engine with scraping and crawling capabilities built with Django, MySQL, and Android app

Adult Income Prediction

Machine learning analysis of census data to predict individual income levels with comprehensive preprocessing and visualization

Deep Learning with PyTorch

Comprehensive Jupyter notebooks showcasing PyTorch for image classification, object detection, NLP, and advanced techniques

Bengali Clinical Machine Translation

Extrinsic evaluation of machine translation quality for low-resource Bengali clinical texts via downstream tasks

Alzheimer's Disease Detection

Deep learning ensemble approach for multi-class Alzheimer's classification from MRI scans

Word Alignment Techniques in Machine Translation

Implementation and comparison of statistical word alignment models for machine translation

Help The Needy

Online donation and volunteer coordination platform for social welfare in Bangladesh