Create searchable video references for data scientists before training

Your searched source videos go untracked, causing dataset gaps and delayed example selection. Data scientists get searchable references and an alert so fresh examples are ready before the next model run.