Saving a fitted Spark ML PipelineModel enables:
Select an answer to reveal the explanation.
Short Explanation
Save PipelineModel → reload identical featurize+predict path later.
Full Explanation
Persistence is required for production batch transforms.