I’m Tom Koeniger, and I have been with Software AG in Darmstadt, Germany, since 2022, when I joined as a bachelor’s student in the research department.
My early work there focused on the data behind AI models. Within German and EU-funded research projects, I evaluated training data for fine-tuning, assessing which data was suitable for improving a model’s performance. I also worked on prompt-engineering techniques for generating new training questions.
I then moved to the Adabas and Natural department. There, I developed a prototype of an MCP server, which was my first substantial experience connecting AI capabilities to established enterprise systems and shaped my current focus on workflow automation.
After completing my bachelor’s degree, I continued as a dual student and am now pursuing my master’s in the same format, combining academic study with practical project work. My current work focuses on AI and workflow automation. I build with the Model Context Protocol (MCP), Agent Skills, and existing infrastructure to speed up processes that are still handled manually and let them run on their own. What I value most is taking an idea and turning it into something that works and can be put to practical use.
Working within the same organization over the course of my studies has helped me settle in properly and find my way around. The continuity is what let me really understand how the organization works and establish myself within it.
