About Me
Mafizur Rahman is a PhD researcher in Electrical Engineering at Prairie View A&M University, specializing in machine learning, neuromorphic computing, and analog deep neural networks (DNNs). At the CREDIT Center, he collaborates with Sandia National Labs to optimize analog DNN implementations, achieving near-digital accuracy, and with USC's Melady Group to enhance the interpretability of transformer-based models. His research spans deep learning optimization, adversarial robustness, and hardware-aware AI acceleration, driving innovations that bridge digital and analog computing for next-generation AI systems.
Academic Journey
Ph.D. in Electrical Engineering
2025 - 2028 (Expected)Prairie View A&M University, Texas, USA
Advisor: Dr. Lijun Qian
Research Focus: Neural network interpretability, analog deep learning, and hardware-aware AI optimization
Working on novel techniques to convert digital neural networks to analog implementations while preserving accuracy and performance. Investigating the impact of hardware constraints on model architecture and learning algorithms.
M.Sc in Computer Science
2023 - 2024Prairie View A&M University, Texas, USA
CGPA: 4.0/4.0 (Highest Distinction)
Thesis: Comparative Analysis of DNN Inference Performance in Analog Accelerators
Conducted extensive research on the performance differences between digital and analog implementations of convolutional neural networks, with a focus on efficiency and accuracy trade-offs.
B.Sc. in Computer Science & Engineering
2016 - 2020East West University, Dhaka, Bangladesh
CGPA: 3.71/4.0
Thesis focused on Bengali sentiment analysis using deep neural networks. Received full tuition fee free merit scholarship for academic excellence in 2019.
Additional Certifications
Fundamentals of Deep Learning
NVIDIA Deep Learning Institute - 2023
Ai for Everyone
Duke University(Coursera) - 2020
