Junhao Zhang is a DPhil candidate in Engineering Science at the University of Oxford, where his research focuses on hardware-accelerated Gene Regulatory Networks (GRNs) for evolutionary machine learning applications.
Previously, he served as a Senior Software Development Engineer at AMD, where he made significant contributions to Flash Attention optimization and spearheaded GPU kernel programming initiatives. His work on framework development has been widely adopted by organizations including Hugging Face and Microsoft.
Zhang's current research explores the intersection of evolutionary algorithms and cellular automata, with particular emphasis on developing fine-grained parallel implementations for GPU acceleration. He holds a Master's degree in Applied Machine Learning from Imperial College London and completed his undergraduate studies in Electronic Engineering at both the University of Liverpool and Xi'an Jiaotong-Liverpool University. His research interests span hardware acceleration, evolutionary computation, and biologically-inspired artificial intelligence systems.
Research Interests
- High-Performance Computing for AI
- Machine Learning
- Bio-Inspired Computing
- Neural Cellular Automata
- Complex Systems Design
Research Projects
Hardware-Accelerated Gene Regulatory Networks for Cellular Automata: Developing a novel fine-grained parallel evolutionary algorithm system to optimize Gene Regulatory Networks for image generation using GPUs.
