Home › GCP ML Engineer › cloudcomputing › mleprototyping › Which Vertex AI feature stores precomputed featu…
Which Vertex AI feature stores precomputed features in a centralised online + offline store — so training + serving consume the SAME feature definitions, eliminating training-serving skew?
AVertex AI Feature Store
BManually computing features at train time and again at serve time
CCloud Memorystore
DCloud DNS
Answer & Solution
Correct answer: A. Vertex AI Feature Store
Feature Store gives a single feature catalog for train + serve (per PMLE §2.1). Manual duplication causes skew; the others aren't feature stores.
Related questions
Which industry-specific Google Cloud ML API should an ML Engineer use to extract structureWhich Google Cloud feature should an ML Engineer use to leverage Apache Spark in a JupyterWhich Vertex AI service should an ML Engineer use to MANAGE labelled datasets (image / texWhich Vertex AI capability should an ML Engineer use to handle PII / PHI during data preprWhich Vertex AI capability should an ML Engineer use to track + compare experiment runs (hWhich Google Cloud service provides COLAB notebooks with enterprise IAM + VPC controls + pWhich Vertex AI service provides managed Jupyter notebooks with GCP integration + idle-shuWhich Vertex AI capability should an ML Engineer use to train custom models on tabular / i