Alec Helbling

I work on diffusion and flow-based generative models — combining theoretical insight (e.g., the role of data geometry in flow matching) with architectural and mechanistic understanding of how these models represent and generate. Most of my work to date has been on models for images, video, and multimodal understanding, and I am increasingly interested in their applications to language and reasoning. I also enjoy developing open-source software libraries and interactive tools around these models.

I have worked with researchers and engineers at IBM Research, the MIT-IBM Watson AI Lab, Georgia Tech, Apple, Adobe, NASA Jet Propulsion Laboratory, Microsoft, and the University of Pittsburgh.