Alert data scientists on pipeline stage changes before model run

Your pipeline stage updates arrive without structured logs, causing model training to use stale labels. Data scientists receive contextual alerts with item details so models are corrected before the next model run.