Code reviews are a central component of collaborative software development, yet they often require considerable manual effort in practice. Ensuring consistent compliance with project-specific standards can be challenging, particularly when these standards require contextual interpretation of code, naming conventions, or documentation.
This paper presents an LLM-based prototype designed to support the application of a documented rule catalog within GitLab merge requests. The agent is integrated into the existing workflow and is triggered automatically by merge request events. It analyzes the source code contained in a merge request and evaluates it against explicitly defined project rules.
To ensure structured and consistent output, a predefined JSON schema guides the model’s response and enables validation before publication. Invalid or non-parseable outputs are detected and not posted to the merge request.
The approach is not intended to replace human reviewers, but to support them in the systematic application of documented project standards. The implementation demonstrates the technical feasibility of integrating rule-based governance with LLM-supported contextual interpretation in a practical GitLab workflow.
Sehyang Na, D. Kostunin· TH Wildau Engineering and Na...· 0 citations
An AI-based approach to supporting control rooms in large-scale infrastructures is presented. Distributed data sources, unclear documentation, and complex system depen-dencies make rapid and reliable decision-making difficult in such environments. The developed assistance system consolidates knowledge from operating manuals, experiential expertise, and real-time data, and makes it accessible through a natural language interface. Technically, the system is based on a locally operated multi-agent architecture that integrates data from moni-toring and control software. A verifiable workflow with fixed feedback loops ensures that inputs and outputs remain traceable and stable. This makes it possible to translate probabilistic mod-els into comprehensible and reproducible action steps—an essential aspect for deployment in safety-critical environments. The development and testing take place in scientific facilities being established at DESY Zeuthen. The paper describes initial results, challenges related to data quality and integration, and the potential contributions of the approach to resilient, transparent, and sustainable AI support systems for control rooms.
D. Kostunin, Julian von Hoerschelmann-Schliwinski· TH Wildau Engineering and Na...· 0 citations