Haojun Qiu

Hi! I am a researcher on the World Models team at Waabi. I received my B.Sc. in Computer Science from the University of Toronto, advised by David B. Lindell and Kyros Kutulakos, where I connected score-based generative modeling with patch-based methods via closed-form denoising, enabling efficient and training-free single-image generative modeling.

In Summer 2024, I worked with Jiajun Wu and Elliott Wu at Stanford University on video score distillation for motion. In Summer 2025, I participated in the Summer Geometry Initiative (SGI) at MIT, where I explored some interesting topics including Monte Carlo geometry processing, log heat diffusion, and fluid simulation.

My current research interests broadly span generative models (diffusion/flow models), video world models, and 3D/4D scene representations.