Md Rezanur Islam
AI & Data Science Engineer
I am an AI and Data Science Engineer specializing in Autonomous Vehicle Safety and Cybersecurity, with a strong foundation in machine learning models, feature selection techniques, and MLOps for the scalable deployment of AI applications. Currently pursuing a PhD in Software Convergence Security, my research focuses on developing data-driven safety and security systems powered by advanced AI technologies. I am now expanding my specialization into Large Language Models (LLMs) to explore their transformative applications across various domains.
Projects
Universal Intrusion Detection System on IVN
UIDS I uses Pearson correlation and ResNet-50 to detect cyber threats. UIDS II improves accuracy with feature engineering, segmentation, and Cognitive Belief-Driven Q-Learning.
Real-Time Aggressive Driving Detection System
MobileNetV1-based system that detects aggressive driving behavior using CAN data, distinguishing driver style from cyber-attacks.
Tech Stack
AI & LLM Development
MLOps & Optimization
Scientific Journals
Adaptive RNN Hyperparameter Tuning for Optimized IDS Across Platforms
Cross-platform optimization using adaptive hyperparameter tuning to improve IDS effectiveness.
CF-AIDS: Comprehensive Frequency-Agnostic IDS on IVN
Frequency-independent IDS approach to enhance threat detection in in-vehicle networks.
CANPerFL: Knowledge-Sharing IDS for In-Vehicle Networks
Federated learning-based IDS using CAN protocol knowledge to enhance detection performance.
LPMSAEF: Lightweight Process Mining Framework for Security Analysis
Process mining-based framework for evaluating software architecture's security and performance.