Priv.-Doz. Dr. Werner Zellinger

I develop foundations for evaluating and certifying machine-learning systems, with publications at NeurIPS, ICML, ICLR, and JMLR. My research focuses on reliability under distribution shifts and uncertainty in real-world deployment.

Since November 2026, I have headed the CD Laboratory for AI Safety and Certification, developing foundations for AI evaluation, robustness assessment, and certification in collaboration with TÜV AUSTRIA.

I also lead the Robust AI group at the LIT AI Lab, Johannes Kepler University Linz, and serve as a Key Researcher at the Linz Center of Mechatronics (LCM), focusing on robust machine learning for digital twins and simulation-based engineering.

Research Areas

• Machine learning under distribution shift
  (out-of-distribution detection, shift detection, shift correction)

• Robustness assessment and certification-readiness
  (test data generation, benchmarks, stress tests, simulation-based testing) 

• Learning theory
  (generalization theory, regularization)

Short BIO

Since 2024, I have led research on robust and reliable machine learning at Johannes Kepler University Linz, supervising PhD researchers and collaborating with industry. Previously, I was a Postdoctoral Researcher at RICAM, Austrian Academy of Sciences (2022–2024), and held research positions such as a Research Team Lead at SCCH GmbH (2012–2022).