My route to AI architecture was indirect. I studied art history, German literature and philosophy. I worked with net art when the internet was still new. That was where I learned to separate theory from practice: writing elegantly about technology does not mean understanding it. That scepticism has stayed with me.

Philosophy brought formal logic. Marketing showed how people absorb, filter and process information. Strategy provided an eye for systems: what connects to what, and where the real problems emerge.

When I began working with AI agents, the issue was obvious: the models are impressive, but without memory, without knowledge structure and without quality control, they are brilliant conversationalists with amnesia. So I started building. Not as a developer, but as someone who understands how knowledge needs to be organised.

The result is an architecture with six memory layers, validation mechanisms, controlled forgetting and a coaching system that recognises work patterns. All local, under my own control, with no cloud dependency.

Beyond that lies a further question: whether AI systems can maintain a model of themselves. Not consciousness, but anticipatory competence. The bridge between cognitive science and AI architecture. That is what I am working on.

At a Glance