Currently, my group is working on two major topics that in one way or another contribute to reliable software: tools for automated program correctness proofs and domain-specific data analysis solutions. Of course, we also use the modern possibilities of distributed systems to solve our problems.

Program Correctness

Our working group is interested in ensuring quality in software through test procedures and (semi-)formal methods. Malte Mues is one of the main developers of the tool GDart for dynamic symbolic execution of Java programs. We are currently working on how to combine dynamic, generally under-approximating methods for software verification with static, generally over-approximating methods. The aim is to find vulnerabilities in Java web applications and other JVM-based programs completely automatically.

The development of GDart began at TU Dortmund University in 2022 and has been continued at the University of Wuppertal since 2025.

Simple Access to Data Analytics for Domain Experts

GeoArmadillo was created as part of the QuARUm project at TU Dortmund University. In extension to the original project, we are working on how a combination of source code generators and suitable libraries can make data analysis pipelines easy to use even for non-computer scientists. Specifically, we are currently looking at case studies from geochemistry and working together with geochemists on domain-specific data analysis solutions and are interested in scaling them to other domains.

We are currently focusing on how the data pipelines genereted can be optimized as automatically as possible in terms of runtime through parallelization. Moreover, we explore how the library must be structured in order to generate correct source code by design that meets the domain problem semantically.