Talks, tutorials and distinctions

A tutorial on the mathematics of neural networks for the Deep Learning Indaba 2026 in Lagos. Starting from the intuition of learning from data, the session builds up cost functions, gradient descent, the artificial neuron and the feedforward pass, then works through backpropagation step by step before a live exploration of how depth, width and activation functions shape decision boundaries. I led the full technical revision of the 101-slide deck, developed with collaborators including Prof. Claire David, including the backpropagation derivations, and adapted worked examples (such as cassava leaf disease classification) to African contexts. Tutorial materials are in the official Indaba 2026 tutorials repository.

Best Paper Award certificate (2nd place, with monetary prize) presented by the conference General Co-Chair and Technical Program Committee Co-Chair for A Novel GPT-2-Based Large Language Model for Expert-Level Radiology Report Generation from 3D Computed Tomography Images, presented orally at the conference and published in the proceedings by Springer Nature (LNICST vol. 697).

Awarded for contributions and research potential in AI and its applications. The fellowship funded six months of research on the InstaGeo team (AI for Social Good), working on physics-informed retrieval of vegetation parameters from Sentinel-2 imagery.

Award of academic excellence as the Salutatorian of the cohort at the AIMS recognition of awards ceremony.

Full scholarship for the AI for Science Master’s programme at AIMS South Africa, funded by Google DeepMind and awarded for outstanding performance, covering tuition, living costs and research computation.