nerdexam
Google

PROFESSIONAL-MACHINE-LEARNING-ENGINEER · Question #292

You work at a large organization that recently decided to move their ML and data workloads to Google Cloud. The data engineering team has exported the structured data to a Cloud Storage bucket in Avro

The correct answer is B. Ingest the Avro files into BigQuery to perform analytics. Use a Dataflow pipeline to create the. To perform analytics, create features, and host features for online prediction from Avro data in Cloud Storage, ingest data into BigQuery, use Dataflow for feature engineering, and store features in Vertex AI Feature Store.

Submitted by ashley.k· Apr 18, 2026Data processing and feature engineering

Question

You work at a large organization that recently decided to move their ML and data workloads to Google Cloud. The data engineering team has exported the structured data to a Cloud Storage bucket in Avro format. You need to propose a workflow that performs analytics, creates features, and hosts the features that your ML models use for online prediction. How should you configure the pipeline?

Options

  • AIngest the Avro files into Cloud Spanner to perform analytics. Use a Dataflow pipeline to create
  • BIngest the Avro files into BigQuery to perform analytics. Use a Dataflow pipeline to create the
  • CIngest the Avro files into Cloud Spanner to perform analytics. Use a Dataflow pipeline to create
  • DIngest the Avro files into BigQuery to perform analytics. Use BigQuery SQL to create features and

How the community answered

(43 responses)
  • A
    14% (6)
  • B
    77% (33)
  • C
    7% (3)
  • D
    2% (1)

Why each option

To perform analytics, create features, and host features for online prediction from Avro data in Cloud Storage, ingest data into BigQuery, use Dataflow for feature engineering, and store features in Vertex AI Feature Store.

AIngest the Avro files into Cloud Spanner to perform analytics. Use a Dataflow pipeline to create

Cloud Spanner is a transactional database, not typically the primary choice for large-scale analytical workloads when BigQuery is available, and it's not optimized for hosting features for online prediction compared to a feature store.

BIngest the Avro files into BigQuery to perform analytics. Use a Dataflow pipeline to create theCorrect

Ingesting structured Avro files into BigQuery is ideal for large-scale analytics. Dataflow is a highly scalable service for complex feature engineering, and Vertex AI Feature Store is specifically designed to host and serve features for low-latency online predictions, ensuring consistency between training and serving.

CIngest the Avro files into Cloud Spanner to perform analytics. Use a Dataflow pipeline to create

Cloud Spanner is not the optimal choice for large-scale analytical workloads from Avro files, making this less efficient than using BigQuery.

DIngest the Avro files into BigQuery to perform analytics. Use BigQuery SQL to create features and

While BigQuery can perform analytics and create features with SQL, hosting features directly from BigQuery for *online prediction* is not suitable due to its analytical nature and higher latency compared to a dedicated feature store.

Concept tested: Data workflow with BigQuery, Dataflow, Feature Store

Source: https://cloud.google.com/vertex-ai/docs/featurestore/overview

Topics

#Google Cloud Services#Data Ingestion#Feature Engineering#Online Feature Serving

Community Discussion

No community discussion yet for this question.

Sign up to join the discussion
Full PROFESSIONAL-MACHINE-LEARNING-ENGINEER Practice