Trustworthy systems for high-stakes work.
We build the products, platforms and research that let organisations prove their software and AI are safe to rely on.
One standard, three ways of working
Most technology companies either sell products, build for clients or do research. We do all three, and each makes the others better.
Our research tells us where AI and software fail people. Our engineering work shows us what real businesses need. Our products turn both into tools anyone can use.
Products
Tools that show, with evidence, that the software and AI you ship can be trusted.
All products- Early access
Wuramark
Continuous security certification for web applications.
Wuramark assesses your application across five layers, grades it from A to F and issues a certificate your customers can verify. The certificate stays valid only while monitoring continues, so it describes your system as it is today, not as it looked on audit day.
About Wuramark - In development
Medalytics
Trust certification for AI in high-stakes settings.
Medalytics will assess AI systems used where mistakes are costly, such as health and education, and certify how they behave with real people, not only how they score on benchmarks.
About Medalytics
Build with us
We design and engineer platforms, data systems and websites for businesses and institutions, with security built in from the first line.
How we workPlatforms and web applications
Customer platforms, portals and booking systems, designed and engineered end to end.
- Product and interface design
- Web and mobile apps
- APIs and integrations
- Hosting and support
AI and data systems
Pipelines, models and analytics that turn your data into decisions, built to be explained and audited.
- Data pipelines and dashboards
- Forecasting and predictive models
- AI assistants and language tools
- Model evaluation and fairness checks
Security engineering
Security reviews and hardening from the team that builds certification tools.
- Application security review
- Secure architecture
- Third-party and OAuth access review
- Incident readiness
Recent work
Labs
Research on how AI behaves in the settings where mistakes are most costly, and how to make it safe there.
Visit LabsCybersecurity AI
AI that defends systems, and defences against AI used to attack them.
Educational AI safety
AI tutors that help students learn without teaching them the wrong things.
Medical AI
Diagnostic and clinical AI that clinicians can check and trust.
Child safety AI
Protecting children in digital spaces without eroding their privacy.
Why we build the way we do
We build in Africa, where systems have to work on unreliable networks, on mid-range phones, under different regulations and against different risks. That is a demanding place to learn engineering.
Software that holds up under those conditions holds up anywhere. Assuming nothing about the network, the device or the user is the same discipline that keeps a system safe in a hospital in London or a school district in Mississippi.
That is the company we are building. We started with security and trustworthy AI because every other system depends on them.
What clients say
The booking platform Afritech built for us transformed how we manage our storage units. What used to take hours of back-and-forth now runs automatically.
Roseberry Mini Storage Afritech Medalytics built our online presence from scratch. The website is clean, professional and has made it so much easier for customers to find and contact us.
Abby's Laundry The technical architecture Afritech delivered for our platform was solid, scalable and delivered on time.
Grain Rain HoldingsGrain Rain Holdings is co-founded by our founder. Working with Afritech on our AI consultancy tools gave us the technical edge we needed.
Data IntellData Intell is co-founded by our founder and our Head of AI Research.