Skip to content

Spotting a problem worth solving with BigQuery ML or AutoML

BigQuery ML lets you build and deploy machine learning models directly on data stored in BigQuery using familiar SQL syntax, ideal for teams with SQL skills who want to avoid data movement. AutoML (via Vertex AI) provides a low-code/no-code interface for training high-quality custom models on structured, image, text, and video data when more advanced model types or greater customization are needed. Choosing between them depends on data location, team skillset, model complexity, and whether the use case fits BigQuery ML's supported algorithms.

Must-know

  • BigQuery ML is best for structured/tabular data already in BigQuery and supports algorithms like linear/logistic regression, k-means clustering, matrix factorization, time-series forecasting (ARIMA_PLUS), boosted trees (XGBoost), and deep neural networks, plus the ability to import TensorFlow models.
  • AutoML (part of Vertex AI) is best when you need to train models on unstructured data (images, text, video) or want Google's automated architecture search and hyperparameter tuning for higher accuracy without writing model code.
  • BigQuery ML keeps data and compute in BigQuery, eliminating ETL/export steps and reducing latency for training and batch prediction on large datasets.
  • Use BigQuery ML for quick iteration and embedding ML directly into analytics/BI workflows via SQL; use AutoML/Vertex AI when you need advanced model customization, explainability, online prediction endpoints, or MLOps pipeline integration.
  • BigQuery ML also supports remote inference by calling pretrained Vertex AI models (e.g., for text generation or embeddings) directly from SQL, blurring the line between the two for certain use cases.
  • For real-time/low-latency online predictions or complex custom architectures beyond BigQuery ML's supported model types, Vertex AI (custom training or AutoML) is the appropriate choice rather than BigQuery ML, which is optimized primarily for batch prediction.
Check this objectiveFree · always available

A retail data analyst has two years of transaction history stored in BigQuery tables. The analyst wants to build a model that predicts which customers are likely to churn next month, using only SQL and without moving the data out of BigQuery. Which approach best meets these requirements?

Your objective map0 tried · 0 right · 41 untouched

What you have tried across GCP ADP's objectives, not a readiness score.

Coverage checked against the published exam guide on Aug 12, 2026.

These are independent practice questions, written against this certification's published exam guide. They are not the certification vendor's own questions, and not the real exam.