I work at the intersection of theoretical linguistics and natural language processing. I am an Assistant Professor in Computational Linguistics (UD1, if you know the Dutch system) at Utrecht University, part of the Language and Communication group, Institute for Language Sciences. I am also a member of the NLP@U special interest group. Before September 2024, I was an Assistant Professor at the University of Groningen, Center for Language and Cognition (CLCG), Computational Linguistics Group.

I teach in the Applied Data Science master program and in the bachelor program Communication and Information Science. I am open to supervise (research) master theses in data science, AI and theoretical linguistics (semantics, mostly).

Are you here to find some of my teaching materials? Probably you are looking for one of these things:

• Two main Colab notebooks from the Transformers course, 2026 edition. Here they are:
  • The miniGPT Colab (co-authored with Menan Velayuthan, where we walk you through an implementation of a small transformer decoder in PyTorch, train it for a bit, and watch it learn;
  • The Instruction-tuning and Agents Colab, showing how to turn a sequence-continuation model into an interactive one, and how to equip it with tools.
• Notes for my ‘Linguistics for Language Technology’ 2023 course at RuG. The course is over a long time ago, but people seem to like the website, so if you are here for the course notes, check them out!

News

• I am going to Tromsø 20-22 October 2026 to give a workshop (seminar? tutorial?) at C-LaBL on using language models in psycholinguistic research.
• Results of an Applied Data Science thesis project that I supervised are in the news: UU press-release, EurekAlert.
• Presented our BabyLM paper in Suzhou.
• Gave a keynote at the SIGTYP workshop at ACL 2025.
• I got an Open Competition XS grant from NWO (Dutch Research Council) – 47k euros to work on something I called “Interacting with AI personae” till March 2027.
• This little book came out at CUP! Preprint is up on LingBuzz.