Hi! I see you've stumbled across my little web page.

I'm a Research Associate in Natural Language Processing at the Institute for Applied Informatics in Leipzig, Germany, as part of the CORAL project. We're building privacy-preserving language models, developing methods of data obfuscation, as well as making models 'forget'. I recently released Grew-TSE, a Python package for the generation of minimal-pair tests (using treebanks) for the evaluation of language-model syntactic performance.

I hail from Dublin, Ireland. I received a BSc Computer Science with Data Science from University College Dublin in 2022 and an MA Linguistics from University Leipzig in 2025, the latter thanks to a DAAD Scholarship. I have also worked on building machine translation models for the EUComMeet project, tools for low-resource languages, and a database for the Institute of Linguistics at Leipzig University. You can find my Master's thesis here, titled "Cross-Linguistic Syntactic Evaluation of Transformers via Treebank Querying".

Feel free to send me an email: gallagher at infai dot org

Recent Publications & Tools

• 2026
  Targeted Syntactic Evaluation of Language Models on Georgian Case Alignment
  Best Paper Nomination Award @ LoResLM Workshop, co-located with EACL 2026
  ACL Anthology · arXiv
• 2026
  text-mallet: A Python Package to Smash Text into Obfuscated Formats
  Tool / Python Package (Work-In-Progress)
  Docs · GitHub
• 2025
  Grew-TSE: Minimal-Pair Test Generation for Syntactic Evaluation
  Tool / Python Package
  Docs · GitHub