Amir Yazdanbakhsh

Georgia Institute of Technology, University of Illinois Urbana-Champaign, Google (United States)

Papers

Total Citations

244

H-Index

About

Amir Yazdanbakhsh is a computer architecture and machine learning systems researcher whose work sits at the intersection of hardware acceleration, approximate computing, and deep learning efficiency. He is perhaps best known for his pioneering contributions to analog computation for general-purpose code acceleration, proposing a circuit-to-compiler framework that leverages limited-precision analog processing to dramatically improve the performance and energy efficiency of modern processors — a body of work that has garnered over 200 citations and reflects his early recognition that conventional transistor scaling was approaching its limits. Yazdanbakhsh has also made significant contributions to the acceleration of emerging deep learning workloads. His GANAX architecture introduced a unified MIMD-SIMD hardware design tailored specifically for Generative Adversarial Networks, addressing the unique computational demands of one of deep learning's most transformative model families. More recently, his DACAPO framework tackles the challenge of continuous learning in resource-constrained autonomous systems, demonstrating his sustained focus on making intelligent systems practical in real-world deployment scenarios such as self-driving vehicles and UAVs. Across his career, Yazdanbakhsh has consistently bridged theoretical innovation and practical system design, making him a noteworthy voice in the ongoing conversation about sustainable, efficient computing architectures.