Embedded Files
On this page, you’ll find the research projects I’ve led, collaborated on, and the core questions that shape my interests as a researcher.
On this page, you’ll find the research projects I’ve led, collaborated on, and the core questions that shape my interests as a researcher.
Supportive AI: an intervention to support group work
Supportive AI: an intervention to support group work
This work has inspired the column Happy Collaboration is not on Our Rubric, written for DUB in March 2026
The problem: Group work in higher education can be brilliant, but also deeply frustrating. Group issues often go unnoticed till it's too late, problems are hard to solve, and decisions about the grade need to be done. Identifying potential problems early can make a lot of difference, but how to do so when groups dynamics are ever-changing, and university classes can be composed of too groups for the teacher to monitor them all?
What I did: I developed and validated an intervention combining digital peer assessment with automated, personalized feedback. Grounded in psychology and educational literature, the framework translates teamwork into measurable aspects and uses algorithms - developed based on human behavior - to help teachers spot groups before problems escalate. Ultimately, it’s a blueprint for human-centric technology: using AI not to replace educators, but to make learning environments more supportive and connected.
Methods: survey design and validation, focus groups, intervies, empirical research and co-design, adaptive algorithm development.
Academic Publications:
Identifying Students' Group Work Problems: Design and Field Studies of a Supportive Peer Assessment (Interacting with Computers, 2024)
Decoding Peer Assessment: An Algorithm to Navigate Group Problems Detection (AIED 2025)
Adapting Supportive Feedback to Personality
Adapting Supportive Feedback to Personality
The problem: One-size-fits-all feedback often misses the mark—an encouraging message that comforts one person can easily feel patronizing or irrelevant to another. This is also true in teamwork, where an misplaced good advice can end up exacerbating problems.
What I did: Two adaptive algorithms that tailors emotional support sentences based on individual personality traits and team performance.
Methods: Empirical studies, statistical analysis, empirical evaluations with target users
Academic Publications:
Adapting emotional support in teams: Productivity, Emotional Stability, and Conscientiousness (Frontiers in Artificial Intelligence, 2025)
Adapting Emotional Support in Teams: Quality of Contribution, Emotional Stability and Conscientiousness (AIED 2024)
Human Attitudes Toward AI and Robots
Human Attitudes Toward AI and Robots
The problem: What people say about AI and robots doesn’t always match how they feel beneath the surface. To design trustworthy systems, we need to look beyond self-reported opinions—and understand how cultural background influences implicit, split-second reactions to machines.
What I did: Collaborated on a cross-cultural study comparing Japanese and Dutch participants to examine how culture shapes both conscious and unconscious attitudes toward robots and avatars across different body designs.
Methods: Experiment design in Opensesame & Jatos platforms, statistical analysis
Academic Publication: A cross-cultural comparison on implicit and explicit attitudes towards artificial agents (International Journal of Social Robotics, 2023)
The Lookit Platform: Biological Motion Study
The Lookit Platform: Biological Motion Study
The problem: Studying infant and toddler cognition traditionally requires bringing families into a lab — expensive, slow, and virtually impossible during a global pandemic.
What I did: led the technical and operational deployment of Lookit at Leiden University, programmed and led an experiment investigating infants' perception of biological motion.
Method: Technical and procedural setup of a new research infrastructure, JS-based programming, video-coding and labeling.
Gesprekstechnieken: A Virtual Patient for Practicing Conversation Skills
Gesprekstechnieken: A Virtual Patient for Practicing Conversation Skills
The problem: Clinical and counseling students need to practice high-stakes, delicate conversations, but real-patient training opportunities are limited, expensive, and stressful for beginners.
What I did: Developed the core interactive features for a virtual patient simulation at Leiden University, to be used as part of work groups, programming avatar animations and character behavior to create realistic, responsive practice scenarios.
Methods: Unity3D C# programming, interactive avatar animation rigging, dialog/behavior scripting, and user-experience optimization for educational VR.
🔗 Read the full case study on SURF Communities.
Page updated
Google Sites
Report abuse