The first engagement was specific and hands-on: prototyping with the advanced AI coding agents, and scaling their database across many nodes. I was embedded with the team and in the code, and I ended it myself, deliberately, when I decided to move toward advisory work.

Later, someone I had worked with there came back with a scoped question. They were moving their Python and TypeScript SDKs onto a shared Rust core, and wanted a judgment on where the boundary between the two should sit. Band had every option to hire hands for that, mine included, and asked for a decision instead: a working session with their engineers, then the analysis, then a written recommendation that carries the reasoning under each decision, so anyone revisiting one six months later can see what it was weighed against, and a plan that keeps both SDKs shipping while the migration happens. One fixed fee for the whole of it. It is the work I do now, and the client who asked for it had already had me the other way.