The Agentic AI and LLM Evaluation Lab investigates large language models and agentic AI as enabling technologies for digital solutions in the energy sector. The lab focuses on how LLM based agents can be evaluated systematically with respect to reasoning quality, robustness, consistency, tool use, and suitability for complex domain specific tasks.
The purpose of the lab is to develop knowledge and methods for assessing which models, orchestration patterns, and interaction strategies are most appropriate for agentic AI applications in energy informatics. This includes the benchmarking of models, the testing of agentic workflows, and the study of how AI agents can interact with data, software tools, digital twins, and human users in realistic settings.
Through this work, the lab contributes scientific and technological foundations for the reliable use of agentic AI in the development of future digital solutions for the green transition of the energy sector.