Balyk, Anatolii (orcid.org/0009-0005-5031-6059) and Oleksiuk, Vasyl (orcid.org/0000-0003-2206-8447) (2025) A unified decision-making framework for LLM selection: bridging industrial and educational applications . Scientific Publishing House of the University of Bielsko-Biała, Bielsko-Biała, Poland, pp. 171-178. ISBN 978-83-67652-52-0
|
Text
978836765252018.pdf - Updated Version Download (688kB) |
Abstract
The integration of Large Language Models (LLMs) into professional fields requires a new generation of specialists capable of making informed engineering decisions about their selection and application. This paper proposes a unified decision-making framework designed to bridge the gap between industrial system evaluation and professional competency development in informatics teacher education. The framework’s core is a quantitative multi-criteria model for assessing LLMs based on performance, cost, privacy, and customizability. We demonstrate its dual applicability through two distinct use cases. The first, in industrial predictive maintenance, shows how the framework guides the selection of an optimal open-source LLM for a mission-critical task. The second, in teacher training, illustrates how the same framework is used as a pedagogical tool to develop systems thinking and AI literacy. The discussion highlights the synergy created by this unified approach, fostering a direct knowledge transfer pipeline between industry needs and educational outcomes.
Downloads
Downloads per month over past year
Actions (login required)
![]() |
View Item |


