A bank is building a BigQuery ML model to detect fraudulent transactions. The dataset is highly imbalanced, with only 1% fraudulent transactions. The data engineer needs to prepare the data to improve model recall for fraud detection while maintaining reasonable precision. They also need to ensure that the serving data is processed consistently. Which approach is most effective?

Ignore the imbalance and train the model on raw data.

Oversample the minority class using SMOTE and randomly split the data.

Undersample the majority class and use a random split.

Use class weights in the model training and apply the same feature preprocessing to serving data.

A healthcare startup wants to build a semantic search feature over clinical notes. They plan to use Vertex AI's text embedding API to generate embeddings for each note and store them in a vector database. The notes vary in length from a few sentences to several pages. What is the best practice for generating embeddings to ensure effective retrieval?

Generate a single embedding for the entire note regardless of length.

Chunk the notes into smaller, overlapping segments and generate an embedding for each chunk.

Use a bag-of-words representation instead of embeddings.

Truncate all notes to a fixed length before generating embeddings.

Google CloudProfessional Data Engineer

Domain 4Objective 2

4.2 Preparing Data for AI and ML PROFESSIONAL-DATA-ENGINEER Practice Questions (Page 3)

Part of the Preparing and using data for analysis domain, which accounts for ~15% of the PROFESSIONAL-DATA-ENGINEER exam. Google Cloud does not publish an official question count, but from its 120-minute exam (~50–80 total, ~8–12 in this domain), expect 3–4 from this objective — we provide 23 practice questions to prepare you well beyond it. (estimate)

23questions here

5free pages

5concepts

~15%of the exam

Questions 11–15

11expert · hard
A bank is building a BigQuery ML model to detect fraudulent transactions. The dataset is highly imbalanced, with only 1% fraudulent transactions. The data engineer needs to prepare the data to improve model recall for fraud detection while maintaining reasonable precision. They also need to ensure that the serving data is processed consistently. Which approach is most effective?
Select an answer first
12foundation · easy
Which BigQuery ML function is used to generate predictions for a single row of data that is provided inline, rather than from a table?
Select an answer first
13application · medium
A healthcare startup wants to build a semantic search feature over clinical notes. They plan to use Vertex AI's text embedding API to generate embeddings for each note and store them in a vector database. The notes vary in length from a few sentences to several pages. What is the best practice for generating embeddings to ensure effective retrieval?
Select an answer first
14foundation · easy
In a RAG pipeline, after generating embeddings for document chunks, where are these embeddings typically indexed to enable efficient retrieval?
Select an answer first
15foundation · easy
In BigQuery ML, which SQL function is commonly used to split a dataset into training and evaluation sets for model training?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to PROFESSIONAL-DATA-ENGINEER

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Google Cloud. “PROFESSIONAL-DATA-ENGINEER” is a trademark of its owner, used for identification only.

4.2 Preparing Data for AI and ML PROFESSIONAL-DATA-ENGINEER Practice Questions (Page 3)

Questions 11–15

Which BigQuery ML function is used to generate predictions for a single row of data that is provided inline, rather than from a table?

In a RAG pipeline, after generating embeddings for document chunks, where are these embeddings typically indexed to enable efficient retrieval?

In BigQuery ML, which SQL function is commonly used to split a dataset into training and evaluation sets for model training?

Keep studying — Preparing and using data for analysis