Dr. Guangzhi Tang (唐广智)

I am an Assistant Professor in the Department of Advanced Computing Sciences at Maastricht University, the Netherlands. My research lies at the intersection of edge AI and neuromorphic computing, aiming to make intelligent systems practical under strict energy, latency, and memory constraints. I lead a team developing cost-effective, brain-inspired computing approaches to meet the growing compute and energy demands of modern AI.

My work centers on spiking, event-driven computing for efficient AI. On the system side, I advance algorithm-hardware co-design for edge and neuromorphic processors. On the algorithmic side, I study efficient machine learning and resource-aware training to improve deployability on edge hardware. These efforts translate into applications such as computer vision, brain-computer interface, robotics, and IoT.

I am the PI of an NWO (Dutch Research Council) funded project exploring alternatives to energy-intensive matrix multiplication by developing brain-inspired, asynchronous, and local computation for deep learning. In education, I coordinate BCS2750 Ubiquitous Computing & IoT, focusing on connected, resource-constrained IoT systems, including RTOS, sensing, networking, embedded AI, and edge/IoT applications, and BCS2710 Image and Video Processing, covering fundamental and deep learning-based methods for visual data analysis.

Before joining academia, I was a researcher at imec. I was a core member at imec advancing the SENECA neuromorphic processor, event-based neural networks, and the corresponding software. I completed my PhD at Rutgers University in the United States, advised by Dr. Konstantinos Michmizos. During my PhD, I bridged robotics and brain science by developing robust, efficient, and adaptive Spiking Neural Networks (SNNs) that address a wide spectrum of robotics challenges on neuromorphic processors.