Beyond the Catalog and Drug discovery with AI

🚀 Join our free Budapest Data Science Meetup!
In our first session, Tamás Németh from DataHub will explore the evolution of data catalogs into "Metadata Control Planes," explaining how authoritative context is the secret sauce for preventing "confident but wrong" AI agents.
Then, Daniel Darabos, CTO of LynxKite will take us into the world of biotech, demonstrating how GPU-accelerated screening is fast-tracking drug discovery by targeting G-protein coupled receptors with unprecedented speed.

📅 Date & venue:

• 26 March 2026, Thursday, from 6 pm
• Shapr3D Office 1054 Budapest, Akadémia u. 6. · Budapest

Please note that the event's official language is English! By attending this event you agree to be photographed. Let us know if you would not like to appear in these pictures. We may use the pictures in social media.

Schedule

17:45 gates open
17:45-18:15 Warmup and chit-chat
18:15-19:30 Talks
19:30 - 21:00 Drinks, snacks, networking

Beyond the Catalog: The Control Plane for Modern Ecosystems and Agentic AI
Németh Tamás, DataHub
Software Engineer @ DataHub. Building Datahub, the open-source metadata catalog. Previously led the data-engineering team at Prezi and worked on a platform to fulfill all of the analytical needs at scale. Tamas is also a committer in the Apache Gobblin project.

Abstract:
Traditional catalogs are no longer sufficient for today’s complex, decentralized landscapes. This session explores their transformation into Metadata Control Planes, the operational “central nervous system” for enterprise information. As AI agents replace humans as primary consumers, this infrastructure becomes the critical bedrock for Context Engineering.
Unlike people, these autonomous systems act directly on the information they find without seeking clarification; they require authoritative context to avoid “confident but wrong” outcomes. We will demonstrate how to repurpose your existing lineage, quality signals, and business glossaries to provide the provenance and semantic clarity these models crave. Discover how to evolve a passive inventory into a high-performance engine for reliable, enterprise-grade AI.

Drug discovery with AI: Targeting G-protein coupled receptors with GPU-accelerated screening

Abstract:
GPCRs are an important but difficult drug target. The receptor we are focusing on, called Mas, has an important role in regulating the cardiovascular system and muscle growth. A natural compound that activates Mas is in phase 3 trials for treating old-age muscle loss (sarcopenia). As it's based on a plant extract, only a limited supply can be produced. The search is on for a highly efficient, safe, synthetic alternative! Join us to learn more about our GPU-accelerated solution for scoring billions of potential drug molecules with a variety of models.

**Daniel Darabos **is the CTO of LynxKite. He has studied at BME and worked at Google and Lynx Analytics previously. He's currently applying his experience with building data science and AI applications to drug discovery for age-related diseases.