Engineering the relationships that matter.

Most difficulty with intelligent systems is not in the model. It is in the relationships around it: which context is available, who may decide what, which tools can cause effects outside the system, and how anyone establishes afterwards what actually happened.

This project collects those concerns into one working model, so that the architecture, the controls, and the operating practices can be discussed as a single subject rather than as separate disciplines.

Technology leaders, architects, engineers, and product teams working on AI-native software and sociotechnical systems — people, agents, data, tools, workflows, decision rights, and feedback.

The introduction is written to be read by anyone making decisions about such systems. The technical detail is available on the framework and patterns pages for the people who will build them.

TUN Systemic Design defines the intended behavior and human experience. TUN Systemic Engineering translates that intent into implementation and operational discipline. One philosophy, two stages of the same work — the Engineering project extends the Design work rather than replacing it.

Systemic means considering connected behavior and responsibility rather than optimizing isolated components. A faster retrieval step, a stronger model, or a cleaner interface each improves one part. Whether the system is dependable depends on how those parts relate: what they permit, what they assume about each other, and what evidence they leave behind.