Articles on quality engineering, AI, and trust, published by Kora, Technext, and TechCabal. Same principle as the talks: written from what I have actually built, broken, and fixed.
Each piece links straight to the publisher. Three themes recur: what AI changes about testing, why trust is engineered rather than declared, and how a QA function actually gets built.
AI is not replacing QA, it is ending the executor model. The case for quality architects who decide what gets tested and why, orchestrate AI agents, and catch the subtle fragility AI-generated code introduces.
An interview on building reliability into systems millions depend on: testing identity at Seamfix, standing up Kora’s QA department from nothing, scaling across fragmented African markets, and why good QA is invisible.
Informal testing does not scale. A seven-step model for designing quality as infrastructure rather than a final gate: framing it as business risk, building capability-driven teams, using AI to augment judgement, and governing releases with data.
AI has made test execution a commodity. What it cannot do is decide which workflows carry the most business risk, so the value moves to the people who design that strategy.
In identity and payments, reliability is the product. Users grant trust when a system works every time, which is why good QA is invisible when it is working.
Positioned as a final gate, QA creates friction. Embedded early, it accelerates delivery by reducing rework and surfacing risk before it becomes expensive.
I write on quality engineering, AI adoption, and trust in fintech infrastructure. Tell me your audience and angle.