Give me the project that’s taking a lot longer than it should.

A carve is a fixed-scope engagement for the project that’s taking too long. Leadership pulls one stalled project out of the roadmap and hands it to me, with cover to work outside the ceremony. I do the due diligence, build it in the team’s own stack, and hand it back running. Corp-to-corp, and I’m gone when the team owns it.

The trigger isn’t an outage. It’s the moment leadership gets an internal estimate it can’t accept: two years of design meetings that produced no software, a dependency on a team with its own roadmap, an approval nobody owns. Most stalled projects are not stalled on the technology.

What I need from you is the cover. The brief that works sounds like this: “Feel free to do whatever you need to get it done. Let me deal with the red tape.” You take the red tape. I take the technical gray zone, outside the planning that’s several weeks out. I’ve written that out as its own piece: the estimate leadership can’t accept is the latency before the build, not the build.

Shape and duration come out of the due diligence, not off a menu. On the review board a team had designed for two years, the carve was two weeks to a working prototype, built from their own screenshots; they had it in prod six weeks after I handed it over, and no support request has reached me since. That exit is the point. I hand back a working system rather than a repo, with the skills and the chat lineage in it, so the next engineer — and the next agent — can read why.

For a director or a VP with one project they can’t wait on and the authority to carve it out. Not staff augmentation by headcount, and not a second opinion on a design nobody has started building. If the thing in the way is the estate itself, that’s path to prod. If it’s down tonight, that’s production recovery.