From 27–31 July 2026, the 6th School on Automated Machine Learning (AutoML) took place in Dortmund, Germany, bringing together foundational concepts, recent research developments, and practical approaches in AutoML.

This year’s edition placed a special focus on Agentic AutoML, reflecting the growing role of foundation models and AI agents in machine learning and data science workflows.

AutoML has become an important part of the toolkit for researchers and developers, helping automate and optimize key stages of the machine learning process. With the rapid development of foundation models, its scope continues to expand, opening new possibilities for more efficient, scalable, and intelligent experimentation.

Over five days, the summer school combined introductory lectures, advanced research talks, and hands-on sessions. Topics included:

• Hyperparameter optimization and model selection

• Global and Bayesian optimization

• Neural architecture search

• Applying and scaling foundation models

• Large language models for efficient experimentation

• AutoML with and for large pre-trained models

• Agentic systems for machine learning and data science workflows

Practical sessions also introduced participants to state-of-the-art AutoML tools and frameworks, including SMAC3 for Bayesian optimization and AutoGluon for automated machine learning.

By combining fundamental concepts with current research and practical applications, the school offered participants a comprehensive introduction to the evolving field of AutoML and its intersection with foundation models and agentic AI.

More information and registration here