James Gardner

University of York

Final-year PhD student at the University of York and Co-Founder of pxld.ai, an AI startup for VFX. My research focuses on inverse rendering, scene representation learning, and world models. I previously held a Research Fellowship at the University of York and interned as a Research Scientist at Toshiba.

An MEng graduate and IET Prize winner, I contribute to the open source project Nerfstudio and have reviewed for NeurIPS, CVPR, ICCV, ECCV, and BMVC. My work is published in top-tier venues including TPAMI, NeurIPS, CVPR, and ECCV.

A description of James' research:

My research explores how AI systems can build richer world models from the limited views available in images and video. I study representations that reason beyond the image plane: what lies outside the camera’s view, how light and visibility shape observations, and how future views can be predicted from partial evidence. This work sits at the intersection of computer vision, machine learning, and computer graphics, aiming to learn models that are aware of space, direction, and scene structure. In the long term, it supports more capable systems for understanding, reconstructing, and editing visual worlds, with applications in immersive media, games, VFX, robotics, and augmented reality.