Before I build private AI around your work, here is the one I hold myself to every day.
If you are an engineer, a CTO, or simply the technically curious kind, a claim of "private AI" is worth exactly nothing without something real behind it. So here is something real: the system I actually run.
Thea coordinates a team of specialised AI agents that plan and build in parallel — reaching for the best model for each task, including frontier cloud models — and treats my data exactly the way I insist on treating yours: it stays mine, whatever model runs it. It is where I prove, on myself first, everything I would put in front of a client.
Privacy isn't about where the model runs — it's about control. Sensitive details are redacted before anything leaves, requests are made unlinkable, and an egress firewall decides exactly what may go where. Frontier-model power, without frontier-model exposure.
A coordinator delegates to specialised workers that research, build and verify in parallel — the difference between a chatbot and a system that does real work.
Every change is audited, version-controlled and reversible. Not a demo held together with tape — software built to a standard I would sign my name to.
Designed against GDPR and Switzerland's nFADP from the first line — because the trajectory is regulated, and privacy that is retro-fitted is privacy that leaks.
I did not learn private AI from a course. I live in it. The same engineering that runs my day is what I bring to your work — scaled to exactly what you need, and no more.
A short, no-obligation conversation. I will tell you honestly where a private AI helps your work — and where it would not.
Start a conversationThea is a personal project, shown here as evidence of practice. Your engagement is always its own, built around your work.