Bahar Asgari

Georgia Institute of Technology, University of Maryland, College Park

Papers

Total Citations

125

H-Index

About

Bahar Asgari is a computer systems researcher whose work sits at the intersection of deep learning, hardware architecture, and edge computing. Her research focuses on optimizing the deployment of deep neural networks (DNNs) on resource-constrained devices, addressing the fundamental tension between the computational demands of modern AI and the limited capabilities of edge hardware. Her most influential work, "Characterizing the Deployment of Deep Neural Networks on Commercial Edge Devices" (2019), has garnered 109 citations and provides a foundational empirical analysis of how DNNs perform on real-world edge platforms — a critical reference for engineers and researchers designing efficient AI pipelines. Building on this, her work on low-communication parallelization strategies advances faster DNN inference for robotics and IoT applications. More recently, Asgari has turned her attention to FPGA acceleration, exploring high-bandwidth memory architectures to optimize sparse matrix-vector multiplication for edge deployments. Together, her contributions form a coherent research vision: making powerful neural network inference practical, efficient, and accessible at the edge — a challenge of growing importance as AI becomes embedded in everyday devices and autonomous systems.