Nadine Chang

prof_pic.jpg

Hi! I am a Senior Research Scientist at NVIDIA, part of the Spatial Intelligence Lab. I currently work on multimodal large language models (MLLMs) — for data selection and generation, the pursuit of hallucination-free and robust MLLMs, and perception tasks more broadly.

I received my Ph.D. from Carnegie Mellon University, where I was advised by Martial Hebert and Michael Tarr. My thesis focused on computer vision and language models that operate under real-world data distributions and applications. Prior to my Ph.D., I completed my Masters at CMU, where I was advised by Abhinav Gupta.

I was honored to be selected as an NSF GRFP fellow in 2018.

news

Jun 24, 2026 ZTRS, our zero-imitation end-to-end autonomous driving approach with trajectory scoring, was accepted to ECCV 2026.
May 05, 2026 Our ICML 2026 position paper, Stop Reactively Patching Your Model Every Time and Start Proactive Test-Driven AI Development, was accepted.
Feb 27, 2026 Two papers accepted to the CVPR 2026 main conference — on multimodal hallucination mitigation and scaling-aware data selection for autonomous driving. A third paper, on scalable prompting for AV video captioning, was accepted to a CVPR 2026 workshop.
Jan 20, 2026 DriveCritic, on context-aware, human-aligned evaluation for autonomous driving with vision-language models, was accepted to ICRA 2026.
Oct 28, 2025 GHOST, our multi-round consistency benchmark for hallucinations, was accepted to WACV 2026.