Fadli Damara

Fadli Damara is a PhD student at Machine Learning Group at TU Berlin and a research associate at the Probabilistic Modeling and Inference Lab at BIFOLD. He received his M.Sc. in Electrical Engineering from TU Berlin, completing his thesis "Learning Semantic-Aware Representations for Device-Edge Co-Inference" at Fraunhofer HHI within the Department of Wireless Communication and Networks, where he also actively contributed to the ongoing 6G research efforts.

His research interests currently lie in generative modeling on discrete data (including language, genomics, and statistical physics systems such as spin glasses), primarily focusing on diffusion samplers for Boltzmann distributions, as well as inference-time methods and scaling laws for diffusion models.