Teaching

What I teach

I teach deep learning, from the fundamentals to applications on real scientific data, including images of the Moon.

Imperial College London

Graduate Teaching Assistant, MSc Deep Learning

I help deliver the MSc Deep Learning course at Imperial College London: running tutorials, answering students' questions and supporting practical exercises on neural network fundamentals, optimisation and modern architectures.

ETH Zürich

Lecturer, Space Data (Master in Space Systems)

I teach the final third of the Space Data course in the Master in Space Systems programme at ETH Zürich. My block introduces convolutional neural networks and U-Net architectures for denoising images of the Moon's permanently shadowed regions (PSRs).

These techniques build on the HORUS framework developed by Ben Moseley and Valentin Bickel, and are key to reliable resource mapping and landing-site planning. Students learn the foundations of deep learning (MLPs, CNNs, U-Nets) and then apply them to clean up real PSR imagery.

Lunar crater with a dark permanently shadowed region — Permanently shadowed regions on the Moon, imaged in near darkness and recovered with machine learning.