Electrical & Computer Engineer, currently finishing an MSc in Artificial Intelligence at KU Leuven.
My background is in systems — C, concurrency, GPU programming — and I have spent the last couple of years moving toward machine learning. What I enjoy most sits between the two: models that have to be made to work on real, awkward data rather than benchmarks that have already been cleaned up.
| Project | What it is |
|---|---|
| Deep Matrix Completion for Gene Prioritization | My master's thesis. Which genes underlie a given disease, cast as matrix completion over an extremely sparse (0.06 % dense), positive-only association matrix. Implements five ways of injecting biological side information across two deep architectures, to work out not whether side information helps but how it has to enter the model before it does. |
| Computer Vision Projects | Three end-to-end projects: face detection and recognition compared across HOG, eigenfaces and deep embeddings; scene segmentation and classification probed with adversarial attacks; and puck tracking in air hockey footage using classical CV alone. |
| Deep Learning | A multi-layer perceptron written from scratch in C with hand-implemented backpropagation, and a cache prefetcher that predicts a program's next memory reference with K-means and an LSTM. |
| Undergraduate Coursework | Archived ECE coursework — CUDA, OpenMP, FUSE filesystems, kernel patches, concurrency primitives. Kept as a record of the ground covered. |
Python · PyTorch · scikit-learn · NumPy · C · CUDA · OpenMP · Java · Verilog