CV

Welcome to my CV page. Here, you can find an overview of my education, teaching experience, supervision activities, and academic qualifications in computer science and artificial intelligence.

Research

My PhD research centers on implicit neural representations for visual understanding. Recent work includes end-to-end implicit neural representations for image classification, high-resolution INR classification, and optical flow upsampling. I am also exploring deformable implicit representations, foveation in implicit neural representations, and how structure in neural weight spaces can support learning.

Teaching and supervision

At TU Delft, I have contributed to several deep learning and computer vision courses. I served as Head TA for CS4240 Deep Learning and DSAIT4005 Machine and Deep Learning, where my responsibilities included exam design, question design, and coordination of PhD and MSc teaching assistants.

I have also given lectures on Vision Transformers as part of TU Delft computer vision courses, including CS4245 Seminar Computer Vision by Deep Learning and DSAIT4125 Computer Vision.

In addition to teaching, I supervise student research projects at both MSc and BSc level. Topics have included implicit neural representations, optical flow, image segmentation, diffusion techniques, implicit depth, and the real-world generalization of optical flow models.

Education

• BSc in Computer Science, Delft University of Technology, Intelligent Data Analysis Track, 2020
• MSc in Computer Science, Delft University of Technology, Artificial Intelligence Track, 2023 Thesis: Transformers for Convex Upsampling of Optical Flow, graded 9.5 / 10
• PhD in Artificial Intelligence, Delft University of Technology, 2023 – current PhD Candidate at the Computer Vision Lab, supervised by Dr. J.C. van Gemert

Teaching

• Planned Head TA for DSAIT4005 Machine and Deep Learning 2026–2027.
• Gave a lecture on Vision Transformers as part of the TU Delft course CS4245 Seminar Computer Vision by Deep Learning.
• Gave lectures on Vision Transformers as part of the TU Delft course DSAIT4125 Computer Vision.
• Head TA for CS4240 Deep Learning 2023–2024.
• Head TA for DSAIT4005 Machine and Deep Learning 2024–2025.
• Head TA for DSAIT4005 Machine and Deep Learning 2025–2026.
• Certified second-degree high school mathematics teacher, obtained through the educational minor during my Bachelor’s degree.

Supervision

• Graduated MSc students:

  • Paul Frolke: Diffusion techniques for optical flow
  • Pascal Benschop: When is depth useful for image segmentation?
  • Joost Luijmes: Processing local implicit neural representations
  • Thomas Streefkerk: Optical flow out-of-frame objects
  • Byran He: Implicit CNNs
  • Alexandra Neagu: Implicit depth
  • Petter Reijalt: Optical flow evaluations

• MSc students currently under supervision:

  • Soumen Sinha: How much data goes into learning transformations?
  • Vladimir Rullens: Deformable implicit representations

• Bachelor supervision:

  • Bachelor group of five students: Do optical flow models generalize to the real world?

Publications

Optical Flow Upsamplers Ignore Details: Neighborhood Attention Transformers for Convex Upsampling

A.S. Gielisse. (2023). Master Thesis. TU Delft Repository.

Color Equivariant Convolutional Networks

arXiv:2310.19368 Attila Lengyel, Ombretta Strafforello, Robert-Jan Bruintjes, Alexander Gielisse, Jan van Gemert

Local Attention Transformers for High-Detail Optical Flow Upsampling

Gielisse, Alexander, Nergis Tömen, and Jan van Gemert. Local Attention Transformers for High-Detail Optical Flow Upsampling. arXiv preprint arXiv:2412.06439 (2024).

ARC: Anchored Representation Clouds for High-Resolution INR Classification

Luijmes, Joost, et al. "ARC: Anchored Representation Clouds for High-Resolution INR Classification." arXiv preprint arXiv:2503.15156 (2025).

End-to-End Implicit Neural Representations for Classification

Gielisse, Alexander, and Jan van Gemert. "End-to-End Implicit Neural Representations for Classification." arXiv preprint arXiv:2503.18123 (2025).