Machine Learning Lab

I am a PhD researcher at the University of Freiburg, supervised by Frank Hutter. My research focuses on the automated design and optimization of LLM-based agentic systems, particularly prompt optimization, multi-agent topology optimization, and credit attribution. I study how feedback on a system’s overall performance can guide improvements to its individual components, with the aim of making agentic systems more capable and efficient.

My work includes CAPO for cost-aware prompt optimization and CANTANTE for optimizing multi-agent systems through contrastive credit attribution. I also contribute to promptolution, an open-source framework for prompt optimization, and am co-developing xmas, a library for multi-agent system inference and optimization.

I worked for five years at Airbus as a Dual Student and Data Scientist, working on applied machine learning projects across different departments, including flight arrival prediction and similarity search for aircraft maintenance. I completed my M.Sc. in Statistics and Data Science at LMU Munich in 2025, specializing in machine learning. My Master’s thesis investigated positional biases in long-context language models and received the department’s Best Master’s Thesis award.

For research projects, writing, and information about student supervision, visit my personal website.