Dr Kuangdai Leng is a research scientist with expertise in physics-informed machine learning, generative AI, and numerical modelling.  

He earned his PhD in computational seismology from the University of Oxford, where he developed **AxiSEM3D**, an efficient tool for seismic wave simulations in realistic Earth settings, leveraging natural spatial sparsity in wavefields. He then worked as a postdoctoral fellow at Yale and Oxford before joining Scientific Computing at the Science and Technology Facilities Council as a data scientist. There, he specialised in deep learning — particularly physics-informed operator learning for PDE-driven problems — and explored the use of generative AI, including large language models and multi-agent systems, in scientific discovery.  

At ERP, Kuangdai’s research focuses on inferring land and soil properties from wavefield data using physics-informed machine learning, knowledge bases, and multi-agent systems. His broader interests include AI-driven approaches to tackling unsolved mathematical problems and advancing scientific discovery through computational intelligence.  

For his latest publications, please visit his Google Scholar profile.