
The Luvo 100
Aaron Bembenek
McKenzie Fellow
University of Melbourne
Australia
Verified & Formal Systems
Why Aaron Bembenek is in the Luvo 100
Aaron helps developers of software-analysis tools express sophisticated checks and run them more efficiently. The Formulog collaboration demonstrated performance improvements across published benchmarks and shared the code, data and instructions behind the findings. This provides researchers and analysis developers with reusable infrastructure for checking software, alongside transparent evidence of its performance.
About
Aaron Bembenek is a McKenzie Fellow at the University of Melbourne, appointed in July 2026 (previously Research Fellow), developing tools for analysing software. Together with Michael Greenberg and Stephen Chong, Aaron created and advanced Formulog, which combines Datalog, automated constraint solving and functional programming to express static analyses. This design enables developers to describe complex checks while drawing on tools that solve logical conditions automatically.
In 2024, Aaron and collaborators published research on accelerating Formulog through eager evaluation, changing the order in which logical rules are explored. Combined with compilation, this approach yielded faster analyses on 20 of 23 benchmarks compared with previously published, hand-tuned implementations. The accompanying software package, reviewed by the OOPSLA’24 Artifact Evaluation Committee, includes code, benchmarks, experimental data, analysis scripts and instructions for building on the findings, helping other researchers examine and reproduce the results.
Career highlights
- McKenzie Fellow at the University of Melbourne, appointed in July 2026 (previously Research Fellow), developing program-analysis tools
- Created and advanced Formulog with Michael Greenberg and Stephen Chong
- Co-authored the 2024 study showing faster Formulog analyses on 20 of 23 benchmarks using compilation and eager evaluation
- Co-released supporting software and experimental materials reviewed by the OOPSLA’24 Artifact Evaluation Committee