Serving Models PROFESSIONAL-MACHINE-LEARNING-ENGINEER Practice Questions (Page 7)

Questions 31–35

1. 31foundation · easy

   Which Google Cloud service is used to deploy models for both batch and online inference with managed infrastructure?

   • AVertex AI Prediction
   • BBigQuery
   • CCloud Storage
   • DCloud Functions
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   Which of the following is an example of postprocessing?

   • AEncoding categorical variables
   • BScaling features to a range
   • CApplying a threshold to a probability to produce a binary class label
   • DSplitting data into train and test sets
   Reveal Solution Discussion Select an answer first
4. 33foundation · easy

   What kind of metadata can be associated with a model in the Model Registry?

   • ATraining data, model architecture, and evaluation metrics
   • BOnly the model's name
   • COnly the model's version number
   • DOnly the model's storage location
   Reveal Solution Discussion Select an answer first
5. 34foundation · easy

   A company needs to generate predictions for a large dataset of customer records every night. The predictions are stored in a table for downstream reporting. Which inference approach should be used?

   • AOnline inference with a low-latency endpoint
   • BBatch inference with a batch prediction job
   • CStreaming inference with a Pub/Sub subscription
   • DInteractive inference with a notebook
   Reveal Solution Discussion Select an answer first
6. 35foundation · easy

   What is the purpose of preprocessing in a serving pipeline?

   • ATo train the model
   • BTo format the model's output for the user
   • CTo monitor model performance
   • DTo transform raw input data into the format expected by the model
   Reveal Solution Discussion Select an answer first