Nullius in verba, take nobody's word for it
A studio built on proof, not promises
Nullius Technologies exists to close the gap between “it should be fine” and “we proved it works,” across QA, AI automation, and data.
Our story
The most dangerous phrase in software is “it should be fine.”
It should be fine that the payment flow handles that edge case. It should be fine after the refactor. It should be fine in production. Should is the word we reach for when we haven't actually checked, and in our experience that gap is where the expensive surprises live.
We learned that the hard way, at the demanding end of the market. We spent years building QA automation inside banking, card processing, multi-currency settlement, SWIFT messaging, where a missed defect isn't a bug ticket, it's money moved wrongly and real regulatory risk. When the cost of “should be fine” is measured in rands and audits, you stop trusting intention and start demanding evidence. That instinct carried into everything else we do: full-stack products, and data science recognised in national competition.
So we named the company after the idea: nullius in verba, take nobody's word for it.
It's the 360-year-old motto of the Royal Society, and it's the cleanest definition of good engineering we know. A passing test suite is evidence. A held-out validation score is evidence. A monitored production metric is evidence. Everything else is a hopeful guess in a confident voice.
Today we're an automation studio working with clients worldwide, founders, small businesses, and enterprises alike, helping them modernise QA, put AI to work inside real processes, and turn the data they already hold into decisions. The size of the team on the other side doesn't change the standard. We don't ask you to take our word for anything. We prove it, then we show you the proof.
Where we've worked
Forged in high-stakes environments
Our standard didn't come from a textbook. It came from years inside the places where a mistake moves money, multiple Tier-1 banks and enterprise engineering teams, where testing isn't a formality and “it should be fine” is never good enough.
Tier-1 banking QA
3+ years of QA automation delivered inside major South African banks, card processing, multi-currency settlement, reconciliation, and SWIFT messaging, under real regulatory pressure.
Enterprise test engineering
Automation frameworks, CI/CD test gates, and deployment pipelines built for large engineering orgs with Playwright, Selenium, Cucumber, RestAssured, Java, Python, and TypeScript.
Full-stack delivery
Years of full-stack and AI product development, from agency builds to complete platforms, across React, Angular, Node.js, TypeScript, and PostgreSQL.
The headline result from that banking work: a critical regression cycle cut by 95%, from a three-person week to a 10-minute automated run, for an estimated R12M in annual saved effort. We bring that same rigour to teams of every size.
What we bring
Four disciplines, one standard of evidence.
QA & test automation
Delivered across South African banking: card processing, multi-currency settlement, SWIFT messaging, web, mobile, and performance.
AI & intelligent automation
Auditable workflows, RPA, and AI integrations built on solid guardrails.
Data science & ML
Recognised in national competition: 3rd of 251 in the Nedbank ML Challenge 2026.
Software development
Full-stack builds and AI products, tested, fast, and handed over clean.
How we work
Our mission is simple: help teams of every size replace guesswork with evidence. Three principles guide every engagement.
Evidence-led
Every recommendation comes with the numbers behind it. We measure outcomes, not effort.
Transparent
You get honest reads, including when something isn’t feasible or worth doing. No hype.
ROI-focused
We start with the highest-value problem and prove the return before scaling the work.
Work with a partner who proves it.
If evidence-led engineering is the kind of partner you want, let's talk about the problem in front of you.
Get in touch