I am a Ph.D. Candidate in the Department of Computer Science and Engineering at the University of South Carolina.
I am advised by Dr. Jianjun Hu, and I work as a Graduate Research Assistant in the Machine Learning and Evolution Laboratory
under his supervision. My field of research is solving materials informatics problems using Deep Learning techniques.
I am born and raised in Bangladesh. I have completed my B.S. in Computer Science and Engineering from Bangladesh University of Engineering and Technology in 2019. My aim is to become an effective and impactful Deep Learning Scientist.
(This website is not optimized for cellphones yet. Please view it in a desktop browser)
As a Ph.D. candidate in Computer Science at the University of South Carolina, my research focuses on developing state-of-the-art (SOTA) deep learning models, particularly graph neural networks (GNNs), transformers, and diffusion models, to solve complex problems in materials informatics. I have applied these methods to complex challenges such as crystal structure prediction (CSP), materials property prediction, and generative modeling of materials. My work on crystal structure prediction is similar to the protein structure prediction problem given amino acid seuqence, recently solved by Deepmind's AlphaFold2.
My current research focuses on developing a diffusion-based CSP model for conditional generation of crystals (both crystal lattice parameters and 3D coordinates of atoms), given their chemical compositions (e.g., NaCl, SrTiO3, Cu4FeGe2S7). I am inspired by successful diffusion model-based approaches in similar problems, such as molecular and protein structure prediction.
