Software Security and Testing Researcher
Research Fellow
University of SzegedInstitute of InformaticsI develop open-source techniques and tools for structured fuzzing, test-case reduction, and reproducible fuzzer evaluation.
BibTeX, OpenAlex, and GitHub metrics · Updated —
I am a Research Fellow at the Institute of Informatics, University of Szeged. My work sits at the intersection of software testing and security: I build automated methods and open-source tools for finding bugs in complex software systems, with a focus on three connected problems — how program structure can guide deeper fuzzing, how failing inputs can be reduced to their essential core, and how fuzzers can be compared fairly and reproducibly. I am the author of Grammarinator, an ANTLR v4 grammar-based fuzzer, and Picire/Picireny, parallel and grammar-aware implementations of delta debugging, alongside newer work such as FuzzMeter for systematic fuzzer benchmarking. I completed my PhD at the University of Szeged in 2019 on fuzz testing and test-case reduction, and my earlier work spans JavaScript engine testing and the security evolution of web browser engines.
I also contribute to the research community by serving on the program committees of the IEEE/ACM International Conference on Automated Software Engineering (ASE) and International Fuzzing Workshop (FUZZING), and by regularly reviewing for Science of Computer Programming (SCP), IEEE Access, and ACM Transactions on Software Engineering and Methodology (TOSEM).
How can program structure guide deeper, more effective testing?
I build grammar-based generators that turn language specifications into effective test inputs, and combine structure-aware generation with coverage-guided feedback to reach code that random mutation misses.
How small can an input become without losing its failure?
I develop parallel and grammar-aware variants of delta debugging that shrink failing inputs to a minimal, readable core — making bug reports faster to triage and easier to understand.
How can fuzzers be compared fairly and reproducibly?
I design methodology and tooling for repeatable fuzzing campaigns — independent measurement, replay, and reporting — so that claimed improvements can be verified rather than taken on trust.
ANTLR v4 grammar-based test generation with Python and high-performance C++ backends.
Systematic fuzzer execution, replay, independent measurement, and interactive reporting.
A configurable and parallel implementation of the classic delta debugging algorithm.
Grammar-aware hierarchical delta debugging built on Picire and ANTLR v4.