Weijian Zheng

Weijian Zheng is an assistant computer scientist in Data Science and Learning Division at Argonne National Laboratory. He earned his PhD in Computer Science from Purdue University in 2022. His research interests include AI4Science and High-performance Computing. 

Weijian Zheng, an Assistant Computer Scientist at Argonne National Laboratory, develops AI-driven frameworks that enable scientists to analyze complex experiments in real time. In high-energy X-ray diffraction and computed tomography experiments, researchers generate massive datasets to understand material structures at the atomic level.

Zheng’s work focuses on building AI-guided feedback systems, such as FastREI for Rapid Event Identification (REI) and InferCT for 3D machine learning, that analyze this data in situ. These tools allow researchers to gain immediate insights and adjust experimental parameters dynamically. He also created the M2ML (Model and Data Management for Machine Learning) framework, which integrates scientific instruments, edge computing, and high-performance computing (HPC). His research also optimizes data movement between facilities using Globus and streaming connections to the Argonne Leadership Computing Facility (ALCF).

Beyond infrastructure, Zheng applies Large Language Models (LLMs) and Vision Transformers to scientific prediction and software understanding. His work on Dabench-LLM explores how generative AI can be used for large-scale scientific software and benchmarking AI accelerators.

These contributions enable ​“closed-loop” experiments that maximize the efficiency of limited beamtime at facilities like the Advanced Photon Source (APS). His leadership in the field is recognized through his role as Sole PI for an LDRD Innovate award and as an Investigator on an AI for Science project.