
The problem
Building and maintaining software involves many kinds of work: development, investigation, testing, and continued improvement. One person can set the direction but cannot personally perform every specialist task across several products at once.
What I built
I built and continue to develop MPDA to coordinate AI contributors working on different parts of software engineering. It brings real development, checking, investigation, and follow-through into a more organized working environment, while I direct product goals and retain responsibility for consequential decisions.
The difference
It has made the organization of different AI-assisted engineering tasks part of my actual product work, rather than only an idea on paper. My wider goal is a capable one-human product organization: one person sets the direction, and an AI team helps carry the work forward.