Oversee AI/ML Model QA/QC (e.g., Configuration Management, Model Performance) PMI-CPMAI Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   What is the primary purpose of configuration management in AI/ML model development?

   • ATo secure the model's API endpoints
   • BTo track changes to model code, data, and parameters so that results can be reproduced
   • CTo monitor the model's performance in production
   • DTo automate the model's hyperparameter tuning
   Reveal Solution Discussion Select an answer first
2. 12application · easy

   A data scientist has developed a new model and wants to deploy it. According to the team's QA process, what must happen before the model can be deployed?

   • AThe model must be tested on the training data to ensure high accuracy
   • BThe model must be deployed to a staging environment for a week
   • CThe model must be approved by the business stakeholder without technical review
   • DThe model must pass a peer review by at least one other team member
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which metric is commonly used to evaluate the performance of a classification model?

   • AR-squared
   • BMean absolute error
   • CF1 score
   • DRoot mean squared error
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A team is developing a fraud detection model. The QA lead notices that the model's precision has increased from 0.85 to 0.92, but recall has dropped from 0.80 to 0.65. The business requirement is to minimize missed fraud cases. What should the QA lead do?

   • AApprove the model because precision is higher
   • BAccept the model and document the trade-off in the final report
   • CReject the model and ask the team to retune to balance precision and recall
   • DAsk the team to increase the decision threshold to improve recall
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   Why is it important for model code to follow coding standards?

   • ATo speed up the model's inference time
   • BTo increase the model's accuracy on the test set
   • CTo improve readability, maintainability, and reduce errors
   • DTo reduce the amount of training data needed
   Reveal Solution Discussion Select an answer first