The overarching goal of my research is to develop novel Generative AI systems to accelerate scientific research and knowledge discovery, especially for biomedical and healthcare domain. To achieve this goal, I have been working on related research problems since early 2018, and have been fortunate enough to have some first-authored publications in some of the highly prestigious venues, including -- RECOMB 2022,  Bioinformatics (OUP),  Briefings in Bioinformatics (OUP),  Journal of Computational Biology.

My undergraduate thesis was in the area of deep learning for computational biology. I worked as a research assistant at BUET with Prof. Bayzid for 3.5 years on several projects, including -- deep learning based gene-tree imputation for improved species tree estimation and phylogenomic analysis (QT-GILD), protein-protein interaction cite prediction (EGRET, Pair-EGRET), protein structure prediction (SAINT, SAINT-Angle). I continued working with him as a remote collaborator. During my undergrad, I also worked as a remote research collaborator with Prof. Dr. Ehsan Hoque (CS, Univ. of Rochester) for almost 1.5 years on interpretable multimodal deep learning and published a co-first authored paper (HirePreter). Since my graduation, I have been mentoring some brilliant and hardworking undergrad students in BUET CSE. 

After my graduation from BUET, I worked in a research collaboration project on Alzheimer's disease prediction and analysis (funded by the Interstellar Initiative Grant), under the supervision of Prof. Bayzid, Prof. Clara Li (Icahn School of Med. at Mount Sinai),  and Dr. Yuichiro Miyaoka (Tokyo Metropolitan Institute of Medical Science, Japan).

At UMD, I mostly worked on neural representation learning and computational imaging with Prof. Metzler. I also worked with Dr. Max Ehrlich (UMD CS; Nvidia AI) on adversarial domain generalization and effective spatio-temporal encoding for human pose estimation and forecasting. 

Please visit my Google Scholar profile to see the list of my completed works.