Focused on building ML systems that are not just powerful, but fair, explainable, and grounded in real clinical data. Research with engineering rigour and real-world purpose.
Active
Current research.
In Progress
Bias Detection & Mitigation in Clinical Healthcare Text
Developing a multi-layer framework for detecting and mitigating implicit and explicit bias in clinical discharge notes using transformer-based NLP models. The system spans corpus construction, supervised classification, explainability auditing, and bias-aware mitigation — designed to improve fairness in AI-assisted clinical decision-making.
BERT-familyLLMsQLoRAFairness in AIExplainabilityHealthcare NLPMIMIC-IIIStatistical Validation
01
Dataset Construction
Complete ✓
02
Human Annotation
Complete ✓
03
Model Training
In Progress
04
Explainability & Mitigation
Upcoming
05
Evaluation & Publication
Upcoming
Published
Publications.
01
arXiv 2024 · Cryptography & Security
A Data-Driven Predictive Analysis on Cyber Security Threats with Key Risk Factors
Fatama Tuz Johora, Md Shahedul Islam Khan, Esrath Kanon, Mohammad Abu Tareq Rony, Md Zubair, Iqbal H. Sarker