Shimeng Yu Pioneers the Future of AI Hardware
By James Moreau | 07 Aug, 2025

Georgia Tech professor Shimeng Yu’s work on semiconductors and integrated circuits makes AI hardware faster, smarter and more efficient.

Groundbreaking research on semiconductor devices and integrated circuits has showcased Dr. Shimeng Yu’s expertise in emerging non-volatile memories for AI hardware utilized in applications like deep learning accelerators, in-memory computing, 3D integration, and hardware security.  His work on probabilistic learning ecosystems won Intel’s Outstanding Researcher Award in 2023.

Yu has developed a roadmap for creating specialized AI chips that perform complex tasks like facial recognition and voice commands with much greater efficiency in his 2018 review “Neuro-inspired computing with emerging nonvolatile memories”.  That efficiency is due to Yu’s technique for processing visual data at the point of capture rather than through a separate processor, leading to faster response times and significant energy savings.

His 2019 publication “Optoelectronic Resistive Random Access Memory for Neuromorphic Vision Sensors” has been applied in the development of computer vision systems for autonomous vehicles and smart surveillance cameras. 

Yu’s research at Georgia Tech’s Laboratory for Emerging Devices and Circuits, where he is a full professor, is supported by the National Science Foundation among other organizations.  In large part due to the commercial potential of Yu’s work, in 2023 the Electrical and Computer Engineering college received a record $81.3 million in sponsored research funding.

A China native Yu earned his bachelor’s degree in microelectronics from Peking University in 2009.  He moved to California to earn his master of sciences and PhD in electrical engineering from Stanford University, respectively, in 2011 and 2013.  Before joining Georgia Tech in 2018 he had been an assistant professor at Arizona State University since 2013. Since August 2024 Dr. Yu has held the Georgia Tech Dean’s Professorship.