We congratulate our former BIBA colleague Hendrik Engbers on the successful completion of his doctoral defense on April 10, 2026. His dissertation, “Ein Meta-Lernverfahren für die vorausschauende Instandhaltung – Automatisierte Auswahl von Algorithmen zur multivariaten Anomaliedetektion,” was supervised by Prof. Michael Freitag of the University of Bremen and Prof. Fazel Ansari of TU Wien.

In his work, Hendrik presents a method that automates the selection of suitable predictive models for predictive maintenance. A central meta-model learns from past tasks and is continuously improved via a closed feedback loop. The method stands out for its practical applicability: it requires only limited training data, processes a wide variety of sensor signals simultaneously, and dynamically adapts to changing operating conditions. The evaluation results show significant improvements in prediction quality with reduced expert and computational effort—a promising step toward industrially scalable AI solutions.