How to configure Vertex AI Workbench for cost-effective ML experimentation?

You are developing a recommendation engine for an online clothing store. The historical customer transaction data is stored in BigQuery and Cloud Storage. You need to perform exploratory data analysis (EDA), preprocessing and model training. You plan to rerun these EDA, preprocessing, and training steps as you experiment with different types of algorithms. You want to minimize the cost and development effort of running these steps as you experiment. How should you configure the environment?

Create a Vertex AI Workbench user-managed notebook using the default VM instance, and use the %%bigquerv magic commands in Jupyter to query the tables.
Create a Vertex AI Workbench managed notebook to browse and query the tables directly from the JupyterLab interface. Source Reference Answer
Create a Vertex AI Workbench user-managed notebook on a Dataproc Hub, and use the %%bigquery magic commands in Jupyter to query the tables.
Create a Vertex AI Workbench managed notebook on a Dataproc cluster, and use the spark-bigquery-connector to access the tables.

Community Votes

83%

17%

83% of anonymous learners picked answer B. Votes are pick records left by other test-takers — they are not the verified answer.

Community Insight

The question tests the selection of the most efficient Vertex AI Workbench environment for iterative ML workflows, where the common trap is over-provisioning resources with user-managed VMs or Dataproc clusters instead of using the fully integrated Managed Notebooks.

Vertex AI Workbench Managed Notebooks provide the most cost-effective and low-effort environment for ML experimentation involving BigQuery and Cloud Storage. The community consensus strongly favors this option for its seamless JupyterLab integration and reduced infrastructure management overhead compared to user-managed or Dataproc solutions.

Choosing Option A (User-managed notebook) is a common mistake because users assume a default VM is cheaper, but it requires manual management and setup, increasing development effort compared to the pre-configured Managed Notebooks.

Community Discussion (8 comments)

b1a8fae 👍 6 Selected: B

"Managed notebooks are usually a good choice if you want to use a notebook for data exploration, analysis, modeling, or as part of an end-to-end data science workflow. Managed notebooks instances let you perform workflow-oriented tasks without leaving the JupyterLab interface. They also have many integrations and features for implementing your data science workflow." vs. "User-managed notebooks can be a good choice for users who require extensive customization or who need a lot of control over their environment." Seems more like the former -> B

AzureDP900 👍 2

B is right because this option allows you to minimize cost and development effort by using a managed notebook in Vertex AI Workbench, which integrates well with BigQuery and Cloud Storage. You can browse and query your data directly within the JupyterLab interface without having to create a separate BigQuery client or use the bq command-line tool.

pinimichele01 👍 1 Selected: B

see b1a8fae

gscharly 👍 1 Selected: A

agree with guilhermebutzke. Also, this option is easier to reuse in multiple experiments

guilhermebutzke 👍 1 Selected: A

My Answer: A A: Default VM instance is the best to minimize the cost, and the command %%bigquery magic is the most easy way to get data from BQ. B: Not necessary JupyerLab interface to run code. The %%bigquerv magic commands is sufficient to get data and run easily queries. C: Dataproc Hub seems overkill and it is more expensive than a default VM instance. C: spark-bigquery-connector unnecessary to get tables in the notebook. better use %%bigquery.

daidai75 👍 1 Selected: B

https://cloud.google.com/bigquery/docs/visualize-jupyter

shadz10 👍 1 Selected: B

https://cloud.google.com/vertex-ai/docs/workbench/notebook-solution#:~:text=For%20users%20who%20have%20specific,user%2Dmanaged%20notebooks%20instance's%20VM.

pikachu007 👍 1 Selected: B

Option A: User-managed notebooks require VM instance management, adding cost and complexity. %%bigquery magic commands are still needed. Option C: Dataproc Hub adds unnecessary cost and complexity for simple BigQuery interactions. Option D: Spark-bigquery-connector adds complexity and overhead compared to the native BigQuery integration in managed notebooks.

Comments & Corrections

No comments yet — spotted an error or have a note? Share it below.

Expert Analysis

Why the Answer Is Correct

Option B is correct because Vertex AI Workbench Managed Notebooks are designed to minimize operational overhead. They provide pre-configured environments with built-in integrations for BigQuery and Cloud Storage, allowing data scientists to browse and query data directly from the JupyterLab interface without managing underlying VM infrastructure. This aligns perfectly with the requirement to minimize cost and development effort during iterative experimentation.

Why the Other Options Are Wrong

Option A is incorrect because user-managed notebooks require manual administration of the VM instance, increasing development effort. Option C and D are incorrect because they involve Dataproc, which introduces unnecessary complexity and cost for simple data exploration and training tasks that do not require distributed Spark processing.

Community Comment Notes

Comments highlight that Managed Notebooks are preferred for workflow-oriented tasks without leaving the interface (Comment 1). Users noted that user-managed notebooks add cost and complexity (Comment 7), while Dataproc is considered "overkill" for this specific use case (Comment 4). The community emphasizes the seamless integration and reduced need for separate client configurations in Option B.

Official Reference

Exam Strategy

When questions ask to minimize "development effort" or "cost" for experimentation, prioritize "Managed" services over "User-managed" ones. Avoid complex distributed computing solutions like Dataproc unless the problem explicitly requires handling massive datasets that exceed single-machine limits.

Related Analysis

Practice All PMLE Questions

Access 65 questions with complete answers and detailed explanations.

View Full PMLE Practice Test →

← Back to PMLE Study Guide