Who builds this
Built by Dieter Stölzel.
Solo engineer and founder in Berlin. I designed, built and operate four production SaaS products — one of them an EU-AI-Act governance platform. Hire agent.berlin and you work directly with the principal who owns the build end to end: no account managers, no offshore hand-off, every deliverable documented to outlast the engagement.
The track record
Twenty years in enterprise IT, including BMW and Oracle. Delivery at a scale where a missed dependency costs real money, where systems outlive the people who built them, and where “it works on my machine” was never an acceptable answer.
That period is the reason a studio of one can quote a production engagement without flinching. The engineering is the easy half; knowing what goes wrong between departments, vendors and deadlines is the half that decides whether anything ships.
Credentials
Certified Project Management Professional (PMP), issued by the Project Management Institute.
It matters most for the AI Delivery Lead engagement, where the deliverable is not code but a programme that reaches production. The AI part of that role is the part everyone talks about; the delivery part is the part that decides the outcome.
What I’ve built
Four production, multi-tenant SaaS products — live software, not a portfolio of mockups. Each one proves a different thing:
- Argumentree — structured argument mapping and collaborative decision-making. Multi-tenant, real-time, and the origin of the reusable engineering core the other three sit on.
- ArgumenTroupe — panels of AI personas that debate a question from every angle. Multi-agent orchestration, in production rather than in a demo.
- AI-Agentree — an AI-governance and EU-AI-Act platform: decision traces, agent identity, the full audit lifecycle, an MCP server and an SDK.
- Argumentree.AI — an AI research engine that turns a question into a sourced, structured decision map.
The third one is the one that matters for this studio. Selling EU-AI-Act engineering is easy to claim; having built, shipped and operated an AI-Act governance platform is the version of that claim you can check. The compliance lifecycle here is software I run, not a guide I read.
How I work
- Build, don’t advise. Every engagement ends in something that runs. A recommendation without a working system is where most AI budgets go to die.
- Fixed scope, agreed up front. No day rates. An open-ended rate points my incentive away from yours.
- Compliance as a by-product of the build, never a retrofit — because reconstructed documentation is slower, weaker and more expensive than documentation produced while the decisions were made.
- One accountable person. You get the principal, in the room, for the whole engagement.
- A hard separation between studio work and my own products. Client work is client work; nothing built for you feeds a product of mine.
What happens if I’m unavailable
The fair question about a studio of one, and it deserves a direct answer rather than reassurance.
Everything is built to be taken over. Source and infrastructure are in your accounts and your name from the start — nothing runs on a credential only I hold. Documentation is a deliverable with the same standard as the code: architecture, decisions, operational runbook, and the compliance artifacts, written for someone who has never spoken to me.
That is also why the deployment tier is your choice (T1 to T4) and why the AI-Act and GDPR documentation is produced during the build rather than at the end. The goal is a system your team — or your next supplier — can run without me. If a specific continuity arrangement matters for your procurement, ask; it is a reasonable thing to want in writing.
Let’s talk about your agent.
Tell me the workflow that’s eating your team’s time. If an agent fits, I’ll show you how — and if it doesn’t, I’ll tell you that instead.
or write to hello@agent.berlin