David Dinkevich

I’m a Machine Learning researcher and PhD student at the Hebrew University of Jerusalem, where I work under the supervision of Prof. Dani Lischinski. My research focuses on generative models for images and video.

My current work focuses on post-training methods that make diffusion-based video models behave reliably for real users: less prompt fragile, more controllable, and consistent across shots. The goal is to make them work on the first try.

I have experience modifying the internals of existing models. My MSc thesis, Story2Board: A Training-Free Approach for Expressive Storyboard Generation, explores adapting attention mechanisms in transformer-based image models to improve character consistency across frames.

My resume is available here.