Short Bio

Florentina Voboril is a PhD student and Project Assistant at the Algorithms and Complexity group at TU Wien. She holds a Master’s degree in Logic and Computation and a Bachelor’s degree in Software and Information Engineering, both from TU Wien, where she also completed the Excellence program “Bachelor with Honors.” Her academic path includes an Erasmus exchange semester at KTH Royal Institute of Technology in Stockholm and a research internship at the Max Planck Institute for Informatics in Saarbrücken.

Beyond academia, she is passionate about science outreach and education. As a long-standing trainer for coding and robotics workshops for children and a role model for the LEA initiative, she is committed to empowering young people, especially girls, to pursue careers in STEM.

Website: https://www.florentina-voboril.com/

PhD Project - Enhancing Constraint Programming with Large Language Models

Supervised by Stefan Szeider

Florentina researches how large language models can help to solve constraint programming problems more efficiently. Her approach is to use the creativity of LLMs to propose so-called streamliners. Streamliners are additional constraints that can reduce the search space and thus improve the speed of solving a problem. However, the answers of an LLM are not always correct. Therefore, part of Florentina’s work is to implement a system to find out which of the proposed streamliners should be selected.

Publications and Conferences

Journal Papers

• F. Voboril, V.P. Ramaswamy, S. Szeider, “Generating streamlining constraints with large language models”, in Journal of Artificial Intelligence Research, Vol. 84 (2025) doi: 10.48550/arXiv.2408.10268

Conference Proceedings

• F. Voboril, V.P. Ramaswamy, S. Szeider, “Balancing Latin Rectangles with LLM-Generated Streamliners”, 31st International Conference on Principles and Practice of Constraint Programming (CP 2025) doi: 10.4230/LIPIcs.CP.2025.36

• F. Voboril, V.P. Ramaswamy, S. Szeider, “StreamLLM: Enhancing Constraint Programming with Large Language Model-Generated Streamliners”, IEEE/ACM 1st International Workshop on Neuro-Symbolic Software Engineering (NSE 2025) doi: 10.1109/NSE66660.2025.00010

Posters (without proceedings)

• F. Voboril, V.P. Ramaswamy, S. Szeider, “Enhancing Constraint Programming with LLM-Generated Streamliners” presented at the BILAI PhD Summer School, Klagenfurt, Austria, July 2025

Presentations

• “Balancing Latin Rectangles with LLM-Generated Streamliners”, CP 2025, Glasgow, Scotland, August 2025

• “Prompt me if you can - a hands-on workshop about LLMs”, BEST Vienna Summer Course, Vienna, Austria, July 2025

• “StreamLLM: Enhancing Constraint Programming with Large Language Model-Generated Streamliners”, NSE 2025, Ottawa, Canada, May 2025