Big Data Conference Europe 2026
November 25th, 2026The 10th Big Data Conference Europe takes place from November 25th to 27th in Vilnius, Lithuania. TNG is represented with four talks.
In his talk, "Large Models, Small Resources: Customizing LLMs," on November 25th, 2026, our colleague Dr. Fabian Klemm provides insights into the work of the TNG AI research team, which published our popular R1T-Chimera model family. He explains our Mixture-of-Experts approach, in which multiple LLMs are combined into a single model, and reports on the successes and setbacks encountered along the way.
Our colleagues Dr. Dennis Schulz and Gregor Ziegltrum will be part of the program on November 26th, 2026, with their talk titled “We Downloaded Our Lives: What Companies Really Know About Us.” In their presentation, they explore the question of what companies actually know about them. To do so, they exercised their right to access information and asked hundreds of companies, from small online shops to large tech conglomerates, for the personal data they had stored. The results range from redacted PDFs to gigabytes of structured data. Using AI, they attempted to reconstruct their own lives from this information - with insights that were at times surprising and very personal. In this talk, they share their experiences, showcase curious and insightful data sets, and demonstrate live how an agent accesses this data. This will highlight just how detailed digital life histories can already be traced today and what questions this raises for each of us.
On November 26th, 2026, our colleague Thomas Endres delivers the keynote, "Skill issue: Teaching robots to play Counter-Strike with imitation learning." He shows how Robot Foundation Models (RFMs) can be trained to play computer games using imitation learning. The talk provides an introduction to RFMs, explains what data they process, and examines why NVIDIA's GR00T model is suitable for this purpose. He then describes the concept of imitation learning, how training data is collected from games, and how the model is iteratively improved. The talk includes live demos and concludes with an interactive one-on-one challenge in which an audience member takes on the GR00T model on stage.
Our colleague Marc Schuh also contributes to the conference program with his talk "Data Lakes - Lessons Learned from Building, Operating and Connecting them to AI" on November 27th, 2026. The session addresses the construction of a central, scalable data lake architecture for enterprises. Drawing on concrete reference architectures and engineering patterns, Marc Schuh explains how to utilize data mesh, Terraform modules, and dbt tests to ensure data quality and ownership. The talk focuses on the critical role of the user interface for governance and demonstrates how such an architecture provides a cost-effective foundation for safe enterprise-wide AI while improving metadata quality.
Further information about the event can be found here.