Researcher at the intersection of Privacy-Preserving Machine Learning, Computer Vision, and Applied AI — building systems that are technically rigorous and ethically grounded.
I am a MS student in Computer Science at Fordham University and a Graduate Assistant at Dept. of Computer and Information Science. My research sits at the intersection of privacy-preserving machine learning, federated systems, novel view synthesis, and visual place recognition.
Currently I am working on membership inference attacks against federated learning models with differential privacy defenses (DP-SGD / Opacus), and on novel view synthesis for Visual Place Recognition using GenWarp and diffusion-based inpainting, submitted to IEEE TCE.
Before Fordham, I was a Lab Instructor and Adjunct Lecturer at North South University and Green University of Bangladesh, where I taught and designed courses across Data Structures, DBMS, and Software Engineering.
Studying membership inference attacks, shadow model training, and differential privacy (DP-SGD / Opacus) as defenses in federated learning settings.
Novel view synthesis for Visual Place Recognition — analyzing how hallucinated pixels from diffusion inpainting affect recognition performance across real-world datasets.
Applying machine learning to real-world security challenges — including IoT intrusion detection, SDN-based threat detection, and resource-efficient multi-class threat classification in constrained environments.
→ Full list on Google Scholar · Download CV ↓
I am open to research collaborations, PhD opportunities, and conversations about ML, privacy, and computer vision. Feel free to reach out.