Teaching philosophy

I see my role as being an active and empathetic guide on a learning journey that the students take by actively doing and practicing. Two commitments sit behind that. The first is that learning is something students do, not something done to them — so my job is to design an environment in which the relevant learning is genuinely worth the effort. The second is empathy: years of one-to-one tutoring and academic-support work trained me to start from where each student actually is, and I try to keep that individual perspective even when I am responsible for a whole room. In practice that translates into constructively aligned courses, an explicit didactical contract, active learning built around think-pair-share, and feedback that runs in both directions — on the students’ learning and on my own teaching.

I am completing the University teacher training programme (UDTU) at DTU, for which the course below served as my main testing ground; it was developed as part of the UDTU requirements.

Courses

Co-developer and co-responsible

02820 Explainable AI and User Experience — Technical University of Denmark 5 ECTS · BSc course · runs every autumn semester (E3A, Tuesdays 8–12), Lyngby campus

A new course I co-developed with Aasa Feragen, Per Bækgaard and Tommy Sonne Alstrøm for the BSc programme in Artificial Intelligence and Data, and one we now co-run as course co-responsibles. The course connects explainable AI methods with the user-experience side of the question: not just what an explanation says about a model, but whether it actually helps a user make better decisions.

Topics include:

  • Feature-attribution methods (LIME, SHAP, gradient-based methods, LRP)
  • Concept-based explainability (e.g. concept activation vectors)
  • Latent-space visualisation (t-SNE, UMAP) and representation alignment (CCA, CKA)
  • Quantitative and qualitative evaluation of explanations — fidelity, stability, plausibility, comprehensibility
  • Human-centred design principles, UI design, heuristic evaluation
  • Designing and running user experiments to assess whether XAI works
  • Data ethics and human alignment

Format: lectures, discussions, hands-on exercises and projects across 13 weeks, two mandatory hand-ins and an oral exam.

Teaching assistant

Technical University of Denmark, DTU Compute

  • Deep Learning (Graduate course) — spring semester 2022
  • Introduction to Machine Learning and Data Mining (Undergraduate course) — spring and autumn semester 2022

Charles University in Prague, Faculty of Social Sciences, Institute of Economic Studies

Thesis supervision

I have co-supervised the following BSc and MSc theses at DTU Compute, together with Professor Lars Kai Hansen. Several led to peer-reviewed publications, which can be found on the publications page.

MSc theses

  • Mathias Gilbert — Explaining AI in realtime (expected 2026)
  • Christian Øster Hyltoft — Uncertainty and Reliability in Large Language Models – An Analysis of Existing Methods and Assumptions (expected 2026)
  • Lasse Møller Sørensen — From Words to Actions: Language-Guided Robot Action Planning with V-JEPA 2-AC: An Empirical Study (2026)
  • Mikkel Godsk Jørgensen — Towards Large Language Models for Educational Technology (2024)
  • Rasmus Ørtoft Aagaard — Towards factual certainty in Natural Language Processing using explainability methods (2023)
  • Ellen Marie Gaunby Jørgensen and Maria Mandrup Fogh — Concept-based Explainability within NLP using Knowledge Graphs (2022)
  • Federico Emiliano Crespo Collazo — Analysing Invariance and Equivariance in Data Augmentation Using Explainability Methods (2022)

BSc theses

  • Albert F. K. Hansen and Marcus Zabell Olssen — Measurement of AI Free Will (expected 2026)
  • Sebastian Ray Mason — Equivariance and interpretability in representation learning (2025)
  • Mikkel Godsk Jørgensen — Improving Visual Classification using Textual Hints (2021)

I also supervise student projects in various machine-learning courses at DTU on a continuous basis.

Academic support

Since 2022 I have worked as an Academic Support Teacher at DTU, helping students with special needs — typically with building an academic overview of their courses, prioritising and planning, developing study techniques, and supporting well-being. This work predates and informs much of my teaching philosophy above.