Herbsttagung 2026
Tucholskystraße 2
10117 Berlin
Thank you very much for your great interest!
The autumn conference of the German Data Science Society brings together practitioners from industry, startups and research for one day in a single track. The focus is on what it takes to bring data science, GenAI and AI agents into production, and on sovereign and trustworthy AI. Most talks are in English.
MORNING
Opening
- Welcome by the GDS board, with Google as host
Opening talk
- „How to deliver a successful Data Science project in business“
Zlata Tanovic HelloFresh - „Soofi: A European LLM. The Next Step: Modular, Self-Improving AI.“
Prof. Dr. Alexander Löser Berliner Hochschule für Technik (BHT)
Data first: quality, collection, tabular foundation models
- „Adaptive Sampling with a Gravity Prior: Can Tabular Foundation Models Better Predict Which Data Should Be Collected?“
Christian Hoops PredictiveStatistics.eu / TU Dortmund - „Automated Data Quality Control in the Age of Generative AI – An Empirical Evaluation of Leading Large Language Models“
Niklas Schüschke Mittelstand-Digital Zentrum Rostock and Johannes Rieper Protiviti - „The Model Was Never the Hard Part: Agents for Predictive Data Science“
Mattia Dimauro Predictabl
Sovereign and trustworthy AI
- „Trustworthy, Reliable & Sovereign AI – When can we Count on AI?“
Prof. Dr. Gerard de Melo Hasso Plattner Institute / University of Potsdam - „Trust, but Trace: Observability for Agents that act“
Sasa Fajkovic Trade Republic
Impulse talk
- „AI trends and European sovereignty“
Jörg Bienert KI Bundesverband (German AI Association)
AFTERNOON
Fireside chat
- „Who Controls AI When AI Takes Control? A Fireside Chat on Healthcare Data, Agentic AI & Digital Sovereignty“
Bettina Goerner Chief Data Officer, Eterno Health, in conversation with Prof. Dr. Anne Schwerk IU International University of Applied Sciences, GDS board
Agents in production: governance, safety, observability
- „When Performance alone is not Enough - a Multidisciplinary View on Clinical Decision Support“
Dr. Roland Roller DFKI (German Research Center for Artificial Intelligence) - „Building AI Systems with Event-driven Operational Agents and Real-time Context Serving“
Steffen Hoellinger Field CTO at Confluent, an IBM Company - „MLOps for Multi-Agent architectures“
Kai Sahling Founder and CEO, &act
Agents at work: lessons from practice
- „Can You Trust What Your Agent Just Told a Customer? A Certificate-Based Approach“
Dr. Lukas Schack, Ulrich Arndt, Michele Zilli TUI - „When agents run the experiments: AI-first model development at Sixt, and what stayed human“
Dr. Vasyl Hafych Sixt - „Agents, Cost, and the Limits of Automation: A Production Case Study“
Zino Lang Avolang - „A Treasure's Impact: AI Agents Turning Hospital Data into Better Staffing and Care“
Dr. Martin Weiss Dataciders
Final Talk
- „Avoiding the AIceberg: How blindly trusting tech hurts us all“
Chris Heilmann VP Developer Relations, WeAreDevelopers
Closing
- Book launch: „Data Science und KI in Theorie und Praxis“
Launch of the new 2nd edition of the Springer Spektrum book
Andreas Gillhuber GDS - Closing and outlook: GDS Board
The registration fee for non-members is €120 (regular) or €35 (students), and for members €60 (regular) or €20 (students).
Together with the invitation to this event, members of the GDS e.V. will receive a discount code which, when entered, allows them to pay only the reduced registration fee of €60 or, in the case of student members, only €20.
Non-members: We would be very pleased to welcome you as a member of the GDS e.V. You can find the corresponding application form here: Become a member. Upon receipt of your application form, you will receive a discount code from us so that you can take advantage of the reduced registration fees intended for members for this event.
For student non-members, the discount code is STUDENT2026HTGDS. Please note that if you use this discount, you must present your student ID at the entrance.
Please also note that the number of participants is limited due to room capacity. Be sure to register early!