Annotation

Annotation: An important practical aspect of the discipline “Software Engineering” is working with version control systems (VCS) and systems for managing dependencies between software packages. To develop skills in this area, teachers have to create various educational tasks. The rapid development of the information technology (IT) sphere leads to the mass character of professional training of IT specialists, which complicates the process of preparing educational and practical problems, and makes the problem of plagiarism of solutions especially acute, as well as the problem of thoughtless, mechanical use of large language models (LLM) for solving educational exercises. In this regard, the most promising approach to the creation of practical problems in the field of software engineering is the development of task generators that allow to automatically obtain an individual version of the educational task for each of the students. The aim of the current work is improving the quality of IT personnel training in the scope of the discipline “Configuration Management”. We propose two new task generators, namely: a task generator for working with git from the command line and a task generator for resolving dependencies of a package. The results show that the use of LLM does not allow students to reliably solve these new generated tasks. The developed task generators are used to assess the knowledge and skills of students during the midterm and final assessments within the framework of the discipline “Configuration Management” of RTU MIREA.