Wires projects 7 through 11 — serving, A/B testing, performance reporting, monitoring, and the retraining pipeline — into one complete, observable MLOps system. Running everything at once costs roughly $5-10 for a two-hour session; tear it all down immediately after with the provided script. Nothing here is new — it is everything else in this path, connected.
Wiring serving, A/B testing, reporting, monitoring, and retraining into one observable system is the single strongest portfolio piece for an ML Engineer or MLOps title — it is the whole job, in one project.