Own entire services across our end-to-end encrypted, security- and privacy-first AI stack, architecting each one from scratch with security built in from the ground up.
Build the private AI layer for confidential work.
PrivateBox is hiring its founding engineers. We care what you’ve built, not where you worked.
A 10-minute screen. Show, don’t tell.
Sovereign AI that runs where the data lives.
Every AI tool asks for the same thing in return: your data, on someone else’s servers. For a hospital, a law firm, a bank or a government, that price is too high.
PrivateBox is the way they don’t have to. An on-premise appliance for organisations that can’t put confidential work in the cloud, plus a confidential-cloud option for everyone else.
On-premise appliance
A ring-fenced box that runs inside their building. Models, knowledge and workflows never cross the perimeter. Built for the organisations that can’t put confidential work in anyone’s cloud.
Confidential cloud
The same controlled environment, hosted: an isolated, governed tenancy for everyone else who wants real AI without surrendering control of the data.
Real parts of the core, in your first six months.
Not a sandbox, not “someday.” This is the actual work waiting for our founding engineers, and you’ll pick up a piece of it in week one.
Flesh out and scale single- and multi-tenant deployments across both cloud and local, working the full stack from infra to backend to CI/CD.
Build feature sets like Layer 7 security for every AI interaction, or async encrypted backups to clients’ own data stores.
Work at the bleeding edge of cryptography paired with AI infrastructure to give clients provable data sovereignty.
No spin.
Here’s the bet.
Early-stage and lean on cash; generous on equity and ownership. You’ll own real parts of the security and AI core from week one. It’s a bet, and we’d rather you knew that now.
R25,000–R37,500 / month. Lean, and we won’t pretend otherwise: enough to do this properly, not top-of-market, and it scales as the company does.
1–5% founding-level equity, vesting over four years with a one-year cliff. Founding-level risk earns founding-level ownership.
We’re building toward a JSE listing in 2030. Your equity is a real stake in that outcome, not a lottery ticket we stay vague about. Ask us the cap-table maths; we’ll show you.
You’ll own real parts of the security and AI core from week one. Not a sandbox.
Technical founders who ship, and a chairman who’s built at scale.
Small team, real pedigree. You’d work directly with the founders every day, with a heavyweight operator who’s done the 0-to-scale journey before in your corner.
Ferdie Pieterse
Most recently CEO of Experian Africa, where he turned a traditional credit bureau into a big-data, analytics and cloud-software business. Experian SA was named a Great Place to Work four years running, and he was recognised as CEO of the Year in the region in 2021. Before that: seven years as COO for Standard Chartered across Africa and the Middle East (25 markets), and an executive director (CFO, then COO) of the JSE-listed Kelly Group. CA(SA), Certified Information Systems Auditor, and an Oxford Saïd Business School executive programme. Across a 33-year career he’s taken businesses from one stage to the next in 40 countries, and he’s chosen to spend this one helping build PrivateBox.
Milan Oerlemans
- Coding since 13; built across fintech, adtech and edtech
- Technical lead at a venture builder spanning ~15 startups
- Shipped Sitetor (AI search) and Binuwe (client-side-encrypted remittance)
- Owns the architecture and security core: encryption, key management, edge AI infra and DevSecOps
Jay van der Walt
- Shipped his first hardware product at 18
- At 16, founded J-Mount, an off-road mount brand now selling internationally
- Owns product and the physical appliance
- Turns the box from a spec into something a client can rack and run themselves
Kian Trathen
- Ran M&A deals at EY in London before coming home to build
- Saw the confidential-AI gap up close: work too sensitive for the cloud
- Owns how PrivateBox reaches legal, finance and healthcare clients
A founding-engineer seat here means building alongside founders who ship, with someone who’s already done 0-to-scale backing the company.
Pre-seed, not pre-traction
A generalist who learns fast and ships.
We don’t care about your degree or your years. We care what you’ve made, and that you believe people shouldn’t have to surrender their data to use AI.
High slope
Your rate of learning matters more than your current altitude. You pick up hard things fast and compound.
Ships, doesn’t stall
You finish things and put them in front of real users. A working artifact beats a perfect plan, every time.
Runs their own infra
You probably run a homelab for fun. Servers, networking, self-hosted everything, because you like understanding the whole stack.
Cares about sovereignty
You think it’s wrong that using AI today means handing your data to someone else. That conviction is the whole company.
Generalist by instinct
Frontend, backend, models, metal: you go wherever the problem is. Founding work doesn’t respect job titles.
Judged on output
No pedigree filter. Show us the repo, the live project, the homelab write-up, and we’ll read what you actually built.
You’ll thrive here if…
- You learn faster than your résumé suggests, and can prove it with what you’ve built.
- You want to own real parts of a product, not a narrow slice of someone else’s.
- Equity and ownership move you more than a big cash number right now.
- You’re energised by ambiguity and an unfinished roadmap.
Probably not, if…
- Top-of-market cash is the thing you’re optimising for today.
- You want a defined scope, a clear ladder, and big-company structure.
- You’d rather be handed the roadmap than help write it.
- You’ve never shipped anything you can actually show us.
How hiring works
Built to respect your time and show you the work. Here’s the whole path, start to finish.
The 10-minute screen
A short form, no résumé. Show us something you’ve built and three lines on why PrivateBox. We acknowledge the same day and screen within 48 hours.
Start the 10-minute screen →Founder conversation (~30 min)
A call with Kian or Jay. The mission, the stage we’re really at, and the honest deal (cash and equity) up front.
Technical conversation (~30 min)
With Milan. Your past work, live systems reasoning, and how you think about AI quality under real hardware constraints.
A short build (a few hours)
A take-home on throwaway infrastructure under NDA (never our client data), plus a short note on how you’d keep it secure.
The founder panel
All four founders, including Ferdie. Real, job-relevant questions and one problem deliberately outside your comfort zone, because we want to see how you think, not what you’ve memorised.
Deep build & decision
Finalists do a 1–2 day build on disposable infrastructure under NDA, never on client data. Then the founders make the final call.
A fast, kind no at every step. For a strong candidate, an offer within two to three weeks, and we won’t leave you sitting in silence.
Show us something you’ve built, and why this matters to you.
A 10-minute screen, not a résumé. A link to something you made, the hardest thing you taught yourself, and why this problem pulls you. That’s it.



