About Me

I work as an applied scientist at AWS AI, focusing on developing high-performing sparse and dense kernels commonly used in Graph Neural Networks (GNN). Our team, DGL (Deep Graph Library) provides the most efficient and scalable GNN solutions in the community.

Before joining AWS, I was a postdoctoral scholar at Berkeley Lab around 2020, where I worked on exascale solutions for microbiome analysis on GPUs and dynamic runtime fusion of parallel operators. In 2019, I completed my Ph.D. at the Ohio State University (a core supporter of our very best Buckeye football team) in CSE under the guidance of Prof. P. (Saday) Sadayappan. My research expertise lies in high-performance computing. Specifically, I specialize in optimizing bandwidth-bound, load-imbalanced sparse kernels for massively parallel architectures such as GPUs and multi-core CPUs. I earned my undergraduate degree in Computer Science & Engineering from the University of Dhaka, Dhaka, Bangladesh.

Latest News

• Feb 2024: Upcoming invited talk at SIAM PP'24, Baltimore, MA.
• Nov 2023: Served at SC'23 as a session chair at Training Graph Neural Networks session.
• Sep 2023: Invited talk at ICIAM'23 (virtual).
• May 2023: Presented paper at IPDPSW'23 on irregular dense operators in GNN models.