Question 90
At a data engineering team preparing AI inputs, an AI project leader faces this situation: A data lake has many similarly named raw folders and analysts cannot identify which is approved for model training. The team must reconcile source quality, permission and readiness evidence. Which principle best explains the underlying issue?
Explanation
Correct Answer
In a data engineering team preparing AI inputs, the underlying issue is: Cloud repositories require dataset catalog, ownership and lineage to distinguish raw from approved sources. It explains the scenario: A data lake has many similarly named raw folders and analysts cannot identify which is approved for model training. The appropriate project response is: Record permitted storage locations, versions and transformation lineage for selected inputs. This can be checked by: A sampled model record traces back to the correct approved raw dataset and revision.
Incorrect Answers
Data requirements must specify meaningful predictors, labels, types and intended use rather than indiscriminate collection. The appropriate correction for the reported case is: Record permitted storage locations, versions and transformation lineage for selected inputs.
Incremental extraction can reduce cost when reliable change-tracking boundaries are available. Fixing that different mechanism would not address the current root cause: cloud repositories require dataset catalog, ownership and lineage to distinguish raw from approved sources.
Business-facing findings must connect numerical deficiencies to real decision and harm implications. Its verification is distinct from the reported condition: A data lake has many similarly named raw folders and analysts cannot identify which is approved for model training. The team must reconcile source quality, permission and readiness evidence.