Question 1

Your team has developed multiple versions of a machine learning model for predicting product demand, and you want to track, organize, and deploy these models effectively. You decide to use Google Cloud’s Model Registry to manage the models. Which practices will help you organize and use the models efficiently? (Select two)

Options :

A : Delete older model versions from the registry after deploying a new version to keep the registry organized and prevent any accidental use of outdated models.

B : Group all models in the registry under a single name without specifying versions to make it easier to find and deploy the latest model.

C : Register models with version numbers and metadata about each model’s hyperparameters, training data, and performance metrics.

D : Use the "Register Model" feature in the Model Registry to add a unique version tag to each model version, indicating its purpose and stage (e.g., "v1.0 - Development" or "v2.0 - Production").

E : Directly upload models to BigQuery to keep model versions alongside the training data for better organization and accessibility.

Answer: C,D

Question 2

You are a data practitioner working for a healthcare organization that stores sensitive patient data in Cloud Storage. To comply with regulatory requirements, you need to restrict access to this data and monitor any changes to permissions. Which access configuration is the most appropriate to ensure restricted access and governance?

Options :

A : Enable Object Versioning to automatically track changes in access controls.

B : Assign the Storage Admin role to all users who need read access.

C : Grant the Viewer role to each user individually on the bucket.

D : Use a group-based IAM policy that grants the Storage Object Viewer role only to a designated group with read permissions.

Answer: D

Question 3

A company wants to migrate a large dataset of historical data from its on-premises storage to Cloud Storage. The dataset is around 100 TB, and the company has a limited network bandwidth, which would make online transfer highly time-consuming and resource-intensive. They need a solution that allows them to quickly and securely transfer this data to Google Cloud without causing disruptions to their ongoing operations. Which data transfer tool should the company choose?

Options :

A : Transfer Appliance

B : BigQuery Data Transfer Service

C : Cloud Data Fusion

D : Storage Transfer Service

Answer: A

Question 4

Which two advantages does using pretrained Google Large Language Models (LLMs) via a remote connection in BigQuery offer when working with natural language processing (NLP) tasks? (Select two)

Options :

A : Google LLMs enable the immediate use of sophisticated NLP capabilities without requiring model development or training expertise.

B : Google LLMs are primarily suited for numerical data processing, making them less ideal for NLP within BigQuery.

C : BigQuery provides built-in support for custom Python packages, allowing advanced NLP operations on Google LLMs directly.

D : Custom layers can be added to pretrained LLMs directly in BigQuery for domain-specific tuning.

E : Integration with BigQuery ML allows users to query and process large datasets using Google LLMs at scale.

Answer: A,E

Question 5

You are building a Looker dashboard for a sales team that wants to track daily and monthly sales performance for different product categories. They want to monitor total sales, average sale price, and the number of items sold. Which of the following features would best allow the team to switch between daily and monthly views on the dashboard without requiring separate charts?

Options :

A : Using a scatter plot with data points labeled by time

B : Adding a date filter with granularity options

C : Applying conditional formatting to highlight trends

D : Using multiple line charts for each time frame

Answer: B

Higher Test Marks with Free Online Associate-Data-Practitioner Exam Practice

Buying Options: