AI Meetup: Kaunas

AI Meetup: Kaunas

Join the Kaunas AI community for inspiring discussions at the AI Meetup: Kaunas. This event brings together researchers, engineers, and industry practitioners - from large-scale alignment techniques to privacy-preserving federated learning and next-generation biomedical signal processing.

📌 Talks & Speakers:

🔹 Bradley-Terry to Constitutional AI: Ranking’s Renaissance

Gediminas Sadaunykas – AI Tech Lead (Centric IT Solutions Lithuania)

In 1952, two statisticians proposed a simple model for comparing sports teams. In 1960, it evolved to power chess ratings. Today, the same mathematical insight underlies the preference-based learning methods behind ChatGPT, Claude, and every major AI model in production. Gediminas explores the surprising lineage of binary preferences, why they are the simplest—and often the most powerful—form of human feedback, and how they became foundational in modern AI alignment and safety.

🔹 Synthesizing a 12-Lead ECG from a Wrist Device Using an Echo State Network

Karolina Eitminavičienė – Junior Researcher (KTU Biomedical Engineering Institute)

Imagine generating a clinically relevant 12-lead ECG from a simple two-lead measurement captured by a wrist-worn device. Karolina introduces a novel echo state network (ESN) approach for synthesizing full ECG data from minimal input, potentially reshaping wearable diagnostics and making cardiac monitoring more accessible than ever.

🔹 Linking Health Data with Trust: Federated Learning in the CVDLINK Project

Rytis Augustauskas – Researcher (KTU Biomedical Engineering Institute & Agmis)

Healthcare data is sensitive, fragmented, and often siloed — making large-scale AI innovation challenging. The CVDLINK project leverages federated learning to enable institutions to collaboratively train models without sharing raw patient data.
Rytis presents the project’s architecture, privacy safeguards, and early results, highlighting how trustworthy AI can transform cardiovascular research and digital health systems.