Technical co-founder for AI startups in Germany and Europe
I am Michael Wutzke, an IT expert for data platforms. An AI startup needs data it can use, a task it can measure and software that stays under control when something fails. In client work I build an information platform with Claude and Codex under machine-checked rules. I consider technical co-founder roles with AI founding teams and work remotely across Germany and Europe.
What I bring as a technical co-founder for an AI startup
-
Data platforms as a foundation
I have built databases and information products, including a global real estate database. For an AI product I start with data origin, quality and rights.
-
AI agents with enforceable limits
On a client's platform Claude and Codex work with developers in one repository. Agents have limited rights, and a person confirms changes to production.
-
A first version and an engineering team
I write code and have hired and led developers. As CIO of a fintech I held the IT budget and led business-critical teams.
-
Founder and funding experience
I founded a company for a global real estate database. KfW and Neuhaus Partners funded it with several million euros. I wrote the plans and pitched the company to investors.
Looking for an AI startup co-founder: what will the first version solve?
An AI startup needs a first user with a specific task. That may be searching a large data set, reviewing recurring cases or handling a step between existing systems. Once the workflow and a useful result are clear, the team can decide which parts a model handles and which parts ordinary software handles.
I begin with the data: Where does it come from, who owns it, how is it updated and how will a user spot a mistake? I then build a first version with real cases. A founding team can test whether the product saves time or solves the task better. That test also gives customers and investors something concrete to examine.
How does a technical co-founder keep AI agents under control?
In current client work I use Claude and Codex on an information platform. Agents and developers work in one repository. An agent’s database role can read but cannot write. Its credentials cannot push to the production branch. A hook checks tool calls before they run. A production change needs a person’s approval outside the chat.
In that project, agents and developers have produced more than 300 RESTful API endpoints and more than 8,000 commits. Those numbers describe one project. For an AI startup I would ask the same questions early: Which data may an agent see, which action may it take and how will the team detect a wrong result?
AI co-founder or developer for an MVP: who owns the decisions?
A developer carries out an agreed task. A technical co-founder decides with the founding team which task to build first, which data foundation will support it and which technical risks must be resolved before a customer meeting. The co-founder builds the first version and makes it possible for other developers to continue.
I have programmed databases and web products, built teams and held a budget and led business-critical staff as CIO of a fintech. My work is in data platforms and software products built with existing AI models. The founding team’s idea needs to fit that field.
Founder experience: which projects worked and which did not?
I have started several ventures. Some succeeded and others did not. A global real estate database I founded in 2004 received several million euros from KfW and Neuhaus Partners. I programmed early products, wrote the business plan and pitched the company to investors. I later started Finance Loop with others, a professional network for emerging technologies in finance.
A founding conversation should also cover assumptions that did not hold. I discuss them in relation to the venture at hand. Funding or a working prototype does not prove that someone will use the product regularly.
Can an AI co-founder work with a team in Germany or elsewhere in Europe?
I live in Frankfurt and work with remote teams across Germany and Europe. For a first call, a description of the user, a few real data examples and the current state of the product are enough. We can then establish who talks to customers, who decides on the product and what the technical co-founder would do personally in the first weeks.
How we assess an AI startup and the co-founder role
-
Define the user's task
We name the first user, the task and a measurable improvement the product could make.
-
Check data and model access
I examine sources, rights, data quality, model access and a set of real test cases.
-
Control the first version
We decide what an agent may read and do, where a person decides and how errors become visible.
-
Agree on collaboration
The founders and their lawyer put responsibilities, time commitments and shares in writing.
Questions companies ask
Do you consider a technical co-founder role in an AI startup?
My work is in data platforms and AI-assisted software development. The user's task, the data foundation and the responsibilities on the founding team also matter for a co-founder role.
Are you an AI researcher or do you build AI products?
My field is data platforms and software products. I use existing models and AI agents in development and check the data, permissions and results.
What experience do you have with AI agents in production work?
On a client's information platform Claude and Codex work with developers. Agents have limited rights, tool calls are checked and production changes need a person's approval.
Have you built a startup with venture investors?
Yes. A global real estate database company I founded received several million euros from KfW and Neuhaus Partners. I wrote its plans, built early products and pitched to investors.
Can an AI founding team outside Frankfurt work with you?
Yes. I live in Frankfurt and work remotely with teams across Germany and Europe. We can meet in person for product or team decisions.
Details on the work behind this page
-
AI agents under the same rules as humans
Limits at the resource, a blocking check before each tool call, and approvals that only a person gives.
-
Building a global real estate database
Managing director and developer of a company funded by KfW and Neuhaus Partners.
-
Projects started and venture funding
What I did myself in the projects I started, what I learned from them, and what became of each of them.
Meet a technical co-founder for an AI startup
In a video call we discuss the first user task, available data and the role on the founding team. We can then assess whether I am the right technical partner.
Searches this page answers
- AI startup technical co-founder
- AI startup co-founder Germany
- technical co-founder AI startup
- AI founding team technical partner
- AI agent startup co-founder
- co-founder for AI product
- artificial intelligence startup founder
- AI product technical partner
- AI agent founder Germany
- AI startup cofounder Frankfurt