A team has deployed a model as a Python UDF and wants to track predictions over time to detect model drift. What is the best way to capture prediction results?

Use the UDF to write logs to a file on the warehouse.

Have the UDF send results to an external monitoring service.

Enable Snowflake's query log to capture UDF outputs.

Create a table that stores the input data, prediction, and timestamp by inserting from a SELECT that calls the UDF.

What is the purpose of a secure UDF in Snowflake?

To allow the UDF to access external data sources

To restrict access to the UDF's source code and protect sensitive logic

To improve the performance of the UDF

To encrypt the data processed by the UDF

A company has a production model deployed as a Python UDF. They have trained a new version and want to test it on a sample of data before switching production. What is the best way to manage this in Snowflake?

Replace the existing UDF with the new version immediately.

Use a stored procedure to switch between versions based on a parameter.

Create a separate database for the new version.

Create a new UDF with a different name for the new version and run both in parallel.

A company has multiple versions of a model deployed as separate UDFs. They want to track which version was used for each prediction to audit model performance. What is the best way to achieve this?

Include the model version as a column in the prediction output table.

Use a single UDF that always uses the latest version.

Store the model version in a separate metadata table.

Use the query history to infer the version.

SnowflakeSnowPro Advanced — Data Scientist

Domain 5Objective 1

Deploy Models in Snowflake SNOWPRO-ADVANCED-DATA-SCIENTIST Practice Questions (Page 5)

Part of the Model Deployment and Operationalization domain, which makes up ~24% of our current practice bank.

27questions here

6free pages

8concepts

Questions 21–25

21application · medium
A team has deployed a model as a Python UDF and wants to track predictions over time to detect model drift. What is the best way to capture prediction results?
Select an answer first
22foundation · easy
What is the purpose of a secure UDF in Snowflake?
Select an answer first
23application · medium
A company has a production model deployed as a Python UDF. They have trained a new version and want to test it on a sample of data before switching production. What is the best way to manage this in Snowflake?
Select an answer first
24foundation · easy
Which Snowflake feature is used to register and manage machine learning models?
Select an answer first
25expert · hard
A company has multiple versions of a model deployed as separate UDFs. They want to track which version was used for each prediction to audit model performance. What is the best way to achieve this?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to SNOWPRO-ADVANCED-DATA-SCIENTIST

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 Snowflake. “SNOWPRO-ADVANCED-DATA-SCIENTIST” is a trademark of its owner, used for identification only.