Md Naim Hassan Saykat

GitHub · LinkedIn · ORCID · Email

I am an Artificial Intelligence researcher specializing in Computer Vision, Medical Imaging, Deep Learning, and Transformer-based architectures.
I am currently pursuing my Master’s in Artificial Intelligence at Université Paris-Saclay, where I develop generalizable, explainable, and efficient deep learning systems for real-world diagnostic and healthcare applications.

My research focuses on building clinically reliable and generalizable AI systems, with an emphasis on cross-dataset validation, explainability, and real-world deployment.

My research interests include:

• Medical Image Analysis (dermatology, radiology, X-ray interpretation)
• Vision Transformers (ViT) and hybrid CNN-Transformer architectures
• Explainable AI (XAI), Grad-CAM, attention mechanisms, feature attribution
• Domain adaptation & cross-dataset generalization (e.g., HAM10000 to ISIC 2019)
• Edge-efficient and deployable AI for resource-constrained clinical settings

I have developed multiple end-to-end research systems, including:

• Generalizable deep ensemble models for skin lesion classification with cross-dataset validation
• Explainable Vision Transformer pipelines for medical image interpretation
• CycleGAN-based unpaired image-to-image translation for medical domain adaptation
• Patent retrieval and re-ranking systems using dense retrieval and cross-encoders
• Transformer-driven explainability frameworks for clinical decision support

My work emphasizes robustness, clinical interpretability, and real-world deployability, with a focus on producing publishable, high-impact research. I am currently preparing manuscripts targeting Q1 journals in medical imaging and AI.

I am actively seeking research opportunities in:

• Research collaborations
• Journal and conference publications
• PhD positions in Computer Vision, Medical Imaging, or Multimodal AI

Academic & Research Links

• GitHub Portfolio
• Skin Lesion Classification Project

If you are interested in collaboration or research discussion, feel free to reach out.