T4Tutorials .PK

NSCT – Model Deployment & MLOps Basics MCQs

1. . Model deployment in machine learning refers to:

(A) Encrypting the model only


(B) Making a trained ML model available for use in real-world applications


(C) Compressing datasets only


(D) Backup only




2. . MLOps stands for:

(A) Compressing ML models


(B) Encrypting ML pipelines


(C) Machine Learning Operations, a practice for deploying, monitoring, and managing ML models in production


(D) Backup only




3. . Continuous Integration (CI) in MLOps involves:

(A) Compressing models


(B) Encrypting code


(C) Automatically testing and integrating changes in code and ML models


(D) Backup only




4. . Continuous Deployment (CD) in MLOps refers to:

(A) Backup only


(B) Encrypting deployment


(C) Compressing deployment


(D) Automatically deploying tested models into production environments




5. . Model versioning is important because:

(A) It tracks changes and allows rollback to previous versions if needed


(B) Encrypting versions


(C) Compressing versions


(D) Backup only




6. . Model serving is:

(A) Encrypting predictions


(B) Making the trained ML model accessible via APIs or web services for predictions


(C) Compressing outputs


(D) Backup only




7. . Monitoring in MLOps helps to:

(A) Track model performance, detect drift, and maintain accuracy in production


(B) Encrypt monitoring data


(C) Compress logs


(D) Backup only




8. . Model drift occurs when:

(A) Backup only


(B) Encrypting drift


(C) Compressing drift


(D) The model's performance degrades over time due to changes in data distribution




9. . A/B testing in ML deployment is used to:

(A) Compress test data


(B) Encrypt test data


(C) Compare performance of two models or variations in production


(D) Backup only




10. . Containerization (e.g., Docker) in ML deployment allows:

(A) Packaging models with dependencies to ensure consistency across environments


(B) Encrypting containers


(C) Compressing containers


(D) Backup only




11. . Orchestration tools like Kubernetes help to:

(A) Compress orchestration


(B) Encrypt orchestration


(C) Automate deployment, scaling, and management of containerized ML models


(D) Backup only




12. . Feature stores in MLOps are used to:

(A) Compress features


(B) Encrypt features


(C) Store, manage, and serve features consistently for training and inference


(D) Backup only




13. . CI/CD pipelines in MLOps help to:

(A) Encrypt pipelines


(B) Automate testing, integration, and deployment of ML models


(C) Compress pipelines


(D) Backup only




14. . Rolling updates in model deployment mean:

(A) Compressing updates


(B) Encrypting updates


(C) Gradually replacing old models with new ones to minimize downtime


(D) Backup only




15. . Canary deployment refers to:

(A) Compressing canary deployment


(B) Encrypting canary models


(C) Deploying a new model to a small subset of users to test performance before full rollout


(D) Backup only




16. . Logging in MLOps is important to:

(A) Backup only


(B) Encrypt logs


(C) Compress logs


(D) Track predictions, errors, and system performance for debugging and auditing




17. . Model reproducibility ensures:

(A) Backup only


(B) Encrypting reproducibility


(C) Compressing models


(D) Same results can be obtained when rerunning experiments with the same data and code




18. . Scalability in ML deployment refers to:

(A) Compressing scalability


(B) Encrypting scalability


(C) Ability to handle increasing data volume and user requests efficiently


(D) Backup only




19. . Cloud platforms like AWS, GCP, and Azure provide:

(A) Compressing cloud ML models


(B) Encrypting cloud ML models


(C) Services for ML model deployment, monitoring, and scaling


(D) Backup only




20. . The main purpose of MLOps and model deployment is to:

(A) Ensure that ML models are reliable, scalable, maintainable, and performant in real-world production environments


(B) Encrypt all ML models


(C) Compress datasets


(D) Backup only




Exit mobile version