Turning fuzzy problems into systems you can trust — and teaching the robots to run them.
"I turn fuzzy problems into systems you can trust — then teach the robots to run them so humans can do the thinking."
I am a Technical Solutions Manager based in Cape Town with 9 years at EDGE Education — a role I did not inherit, I built the case for it by showing exactly which data was missing and what the right workflows would unlock.
I am self-taught across most of what I do. No formal handovers, no playbooks — just curiosity, persistence, and a trail of working systems. In more than a few cases I learned it, built it, and then taught the team.
I design workflows and data models that make time and cost visible, reporting that runs itself, and SOPs that outlive the meeting. Happiest when a 45-minute job becomes a 45-second click. Based in Cape Town and remote-friendly.
Automate first — If I do it twice, I script it. If it needs one more tweak, I build a pipeline.
Make it explainable — Clean structures, clear joins, obvious ownership, plain-English docs.
Design for audit — Log inputs and outputs, validate transformations, version queries.
Calm beats clever — Stable, predictable delivery over fragile one-offs.
Impact first — Start at the bottleneck, fix the source of pain, create lasting capacity.
I treat AI as a tool — like any other in the toolkit, it's only as good as the person using it. I use it for scaffolding, edge-case hunting, and first drafts. I never share sensitive data, I never connect AI directly to my systems, and I always verify before anything ships.
Pressure-test ideas, surface counter-examples, and produce clear summaries I can share with stakeholders.
Draft SQL and Apps Script snippets, scaffold data models, generate first-pass SOPs — then I refine and own what goes out.
Sensitive inputs get redacted. Output gets reviewed. If I can't explain what it built, I don't ship it. The accountability stays with me.
"AI is a tool, not a replacement. The judgment, the context, and the final call always stay with the human."
Never share sensitive data — redact or synthesise before passing anything to an AI tool.
Always own the output — AI gets you to a draft faster, but everything gets reviewed and refined before it's mine to stand behind.
Cite the tool — when AI contributes to a deliverable, I note it. Transparency builds trust.