Manuel Welte is a research associate at BIFOLD and a PhD candidate at TU Berlin, specializing in Explainable Machine Learning. His current research focuses on transforming existing neural network architectures into powerful yet inherently interpretable machine learning models. Before joining the Machine Learning Group, Manuel completed his M.Sc. in Computer Science at Freie Universität Berlin, where he wrote his thesis on pruning Clever-Hans features learned by neural networks.