Public managers commission AI systems they did not build. They procure it, oversee it, and defend it. They were rarely trained to govern it. I research that gap. I write for the people standing in it.
Currently
Since mid-2022 I have worked as a contractor for the Western Cape Government in South Africa. I contributed to its data governance policy framework for AI. I now conduct research for, and co-create the provincial data strategy, run ad hoc labour market analyses, and advise on strategic AI questions as they arise.
The advisory work has one recurring shape: a manager must weigh expert claims about AI without being an expert. My authorship is about making that judgement possible.
The primer series
That work produced a series: the AI commissioning & oversight primers. Each one offers evidence-based, operational guidance to managers who commission AI rather than build or fund it.
The primers’ evidence foundation is fixed. Twenty-five documents. Government audits, peer-reviewed articles, OECD analysis, conference proceedings, and a systematic literature review. No source outside that foundation is cited. Every claim is checked against it.
The evidence is global. The vantage is a working government in the Global South.
How the primers were made
I produced the series with a large language model, Claude Opus by Anthropic. The model searched, retrieved, and synthesised the evidence. I set the questions. I verified the claims. I own the conclusions. The result is not generic output from a prompt. It is structured analysis, grounded in named sources, written for a defined reader.
How I work
Fifteen years as a data analyst and researcher. Quantitative methods in Stata and Python. Disciplined use of language models for structured RAG literature analysis.
I also taught: university entry-level macroeconomics, economic policy and Excel. That habit holds. Take a hard idea and make it usable by someone who has to act on it.
The boundary
I am not a machine learning engineer. I do not build the systems I write about. That boundary is the point. I share the vantage of the managers I write for, and I bring the research training they were never required to have.