Weicheng He

Hi! My name is Weicheng. I am currently a Ph.D. Researcher in Computer Vision and Trustworthy AI at Verimag, CNRS and Université Grenoble Alpes. My research centers onruntime monitoring and AI safety, aiming to understand when deployed vision models fail, detect risks in real time, and reduce unreliable predictions.

I am deeply interested in building safer visual systems. My recent work develops model-agnostic monitoring and mitigation methods across YOLO, Faster R-CNN, and Vision Transformers, reducing hallucination-driven false positives by more than90%.

Before beginning my Ph.D., I worked as an Applied AI R&D Engineer at Verimag on a European research project with Siemens, integrating AI components into verifiable co-simulation workflows. Earlier, I interned at Huawei's Fermat Lab in Grenoble, extending formal simulation frameworks for AI-enabled hybrid systems.