My moat as a Product Engineer
For product teams
I figured out for myself how to collaborate with AI rather than fight it for expertise, and this is what makes me a great Product Engineer. It's my moat.
The Product Engineer role already combines fluency in product with the technical implementation of the planned direction. AI just amplifies it: if the foundation is laid well, the technical implementation can be delegated to AI agents. After developing products for nearly a decade, it was time to put my expertise into a running tool, free my hands from repetitive tasks and give more time to product insights. That's why this app exists.
ADW is an AI Developer Workflow that covers the software development lifecycle (SDLC). Its phases are plan, implement, test, review and document. Plan creates a spec, implement implements the spec, test runs and creates tests, review checks the implementation against the spec, and document writes the docs, so the next ADW run already reads the up-to-date docs where they apply. A run is triggered by creating a GitHub issue and labelling it. I can work on multiple projects at a time by creating several issues per project, but the issues within one project are implemented one after another. Call it single-threaded per project.
What does this give me? While I plan a feature and break it down into atoms, the previously planned feature is running at the same time. I can plan before going to sleep and wake up to a ready feature. And other similar utopian scenarios.
Is it perfect? It's not. Like every product, it wants new features, improvements and fixes, but those come from real usage, and this is what I love as a Product Engineer: bouncing against the real world, not against perfect scenarios in my head of never-shipped products and features.