DevOps and MLOps DY0-001 Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   Which of the following is a key consideration when deploying a machine learning model to a production environment?

   • AThe model must be deployed to the same server that hosts the training environment
   • BThe model's inference latency and throughput must meet the application's performance requirements
   • CThe model must be retrained on every prediction request to ensure accuracy
   • DThe model's training data must be stored in the same database as the production data
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A data science team develops a churn-prediction model in Jupyter notebooks. They manually copy the trained model file to a shared drive, and the operations team manually runs a script to deploy it to the production scoring endpoint. The team wants to reduce errors caused by manual handoffs and ensure the exact same code version that passed validation is what gets deployed. Which practice should they implement first?

   • ASet up a cron job that periodically retrains the model on new data and overwrites the existing model file in the shared drive.
   • BProvide the operations team with direct access to the Jupyter notebook server so they can run the training code themselves.
   • CCreate a CI pipeline that runs automated tests and packages the model artifact with a version tag whenever code is merged to the main branch.
   • DAsk the operations team to document the deployment steps in a runbook so they can follow the same procedure every time.
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   What is the primary benefit of continuous deployment (CD) in a software delivery pipeline?

   • AIt allows new features and fixes to reach users automatically after passing automated tests
   • BIt ensures that all code changes are reviewed by a human before they are released
   • CIt requires a manual approval step before every production release
   • DIt eliminates the need for automated testing in the deployment pipeline
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A team is deploying a model that predicts customer churn. The model is deployed as a batch job that runs nightly. The team wants to monitor the model's performance. Which metric is most appropriate to track for a batch prediction job?

   • AThe model's prediction latency, measured in milliseconds.
   • BThe CPU utilization of the server running the batch job.
   • CThe number of model retraining jobs that fail.
   • DThe model's accuracy on a rolling window of labeled production data.
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A company has a mature CI pipeline for its ML models. The pipeline runs automated tests, builds a container image, and pushes it to a registry. The team wants to implement continuous deployment to a Kubernetes cluster. What is the most important additional component they need to add to their pipeline?

   • AA new container registry to store the images.
   • BA more comprehensive suite of unit tests.
   • CA deployment stage that automatically updates the Kubernetes deployment with the new image and monitors the rollout.
   • DA manual approval step before the image is deployed to production.
   Reveal Solution Discussion Select an answer first