About

I'm an economist who ended up deep in machine learning, or an ML engineer who never stopped being an economist — depending on the day. My PhD (University of Pretoria, 2024) applied natural language processing to central bank communication: teaching transformers to read two decades of SARB policy statements for sentiment, topics, and consistency, and linking that language to how markets react.

These days I work at the South African Reserve Bank, where I'm helping stand up the Research Data Centre and building AI systems — finetuning LLMs for economic sentiment, training computer vision models, and engineering agentic workflows. I'm also a Research Fellow at UP Macrolab, the macroeconomics lab around the SARB Chair in Monetary Economics.

The common thread: most of what economists want to know is locked in unstructured data — text, images, sprawling administrative records. I've worked across all of them, from the SARS tax microdata at UNU-WIDER to health claims at Discovery to MPC statements in my own research.

Off the clock: gaming (Baldur's Gate 3 and Helldivers 2, at last count), padel, golf, Muay Thai, and a lot of running. I cook — the gourmet boerewors roll is a house speciality — and 3D-print accessories for my D&D campaign, which is exactly as nerdy as it sounds.