Head of Research Group: Dr. Gergely LUKÁCS
Members of the Group: Kornél GÁTMEZEI, Tamás Béla KIS, Boldizsár KOVÁCS, Benedek NAVRATYIL, Dr. Zoltán MÁTHÉ, Zita RADVÁNYI, Klaudia SZILI
Contact: lukacs@itk.ppke.hu
Our research group focuses on application-driven research in data-intensive systems and artificial intelligence, combining broad expertise with practical impact. Our work spans relational, horizontally scalable, time-series, unstructured, and vector databases. We study their design, application, benchmarking, and emerging industry trends. In machine learning, our primary focus is on structured data analytics and the application of large language models. Over the past several years, we have benchmarked relational databases, big data platforms, and vector databases . Our current research centers on the evaluation of generative AI systems, including retrieval-augmented generation (RAG), LLM-as-a-Judge, vector embeddings, anonymization, and text-to-SQL.
We see significant societal impact in environmental sustainability and, even more importantly, in strengthening human relationships and communities. We have extensive experience in developing integrated environmental and water management databases. Our interdisciplinary collaborations include the development and evaluation of voicebased social media platforms, social network analysis of hospital chaplaincy communities, and benchmarking the quality of large language model responses to relationship-related questions.
Impact of anonymization on vector embeddings in ITOps data, measured by changes in embedding-based similarity metrics. Kornél Gátmezei, RAG-Enhanced LLMs and Their Evaluation, Master’s Thesis, 2026.
Future research directions, collaboration opportunities
We welcome research, industrial, and interdisciplinary collaborations in data engineering, artificial intelligence, and AI evaluation. We support industry-driven student projects, provide expert consulting, and seek partnerships with social scientists to explore the impact of digitalization and AI on human relationships and communities.
Key publications
- Yang, Z. Gy., Stajer, L. A., Lukács, G. (2025). Under the hood: An inside look at PULI models In: Tómács, Tibor (szerk.) PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON FORMAL METHODS AND FOUNDATIONS OF ARTIFICIAL INTELLIGENCE EGER, MAGYA-RORSZÁG: ESZTERHÁZY KÁROLY KATOLIKUS EGYETEM LÍCEUM KIADÓ
- Horváth, A., Oláh, A., Pintér, A., Siklósi, B., Lukács, G., Reguly, I. Z., Tornai, K., Zsedrovits, T., Máthé, Z. (2025). Anomaly Detection Algorithms for Real-Time Log Data Analysis at Scale. IEEE ACCESS 13 pp . 136288-136311, 24 p.
- Lukács, G., Jani, M. (2016). Analyzing speech and music blocks in radio channels: Les-sons learned for playlist generation. IN: ROBLES, R; TALLON-BALLESTEROS, A J; PIT, PICHAPPAN (SZERK .) ELEVENTH INTERNATIONAL CONFERENCE ON DIGITAL INFOR-MATION MANAGEMENT (ICDIM) NEW YORK, AMERIKAI EGYESÜLT ÁLLAMOK: INSTI-TUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS (IEEE)
