David Holzmüller

David Holzmüller is a researcher at NVIDIA working in machine learning. His research has focused on tabular data, including methods such as TabICL and RealMLP and the TabArena benchmark; he has also worked on uncertainty quantification and active learning.

Career

Cosigned byLoading...

CosignedSign in to view

Elsewhere

NVIDIA

Nov 2026 — Present

Researcher

University of Stuttgart

M.Sc. Computer Science

University of Stuttgart

B.Sc. Mathematics

University of Stuttgart

B.Sc. Computer Science

University of Stuttgart

PhD

Products & media

podcasts.apple.com

Which Tabular Model Should You Actually Use?

A conversation about choosing among tabular foundation models, boosted trees and neural networks, with discussion of TabArena, TabICL and RealMLP.

Media appearance

dholzmueller.github.io