1. What is MLOps ? Junior A. running a candidate model against the incumbent on live data to decide promotion B. recording each training run's params, metrics and artifacts so results are comparable and reproducible C. the ability to recreate a model exactly by pinning data, code, environment and random seeds D. applying DevOps practices to ML — automating the path from data to a monitored production model
2. Which term means: "applying DevOps practices to ML — automating the path from data to a monitored production model"? Junior A. MLOps B. Model artifact C. Train-serve parity D. Reproducibility
3. Which statement is correct? Junior A. MLOps — recording each training run's params, metrics and artifacts so results are comparable and reproducible B. MLOps — ensuring the same transformation code runs in training and serving to avoid skew C. MLOps — reverting an inference endpoint to a previous model version when the new one underperforms D. MLOps — applying DevOps practices to ML — automating the path from data to a monitored production model
4. What is Experiment tracking ? Junior A. reverting an inference endpoint to a previous model version when the new one underperforms B. running a candidate model against the incumbent on live data to decide promotion C. recording each training run's params, metrics and artifacts so results are comparable and reproducible D. applying DevOps practices to ML — automating the path from data to a monitored production model
5. Which term means: "recording each training run's params, metrics and artifacts so results are comparable and reproducible"? Junior A. Experiment tracking B. Pipeline orchestration C. Model artifact D. Train-serve parity
6. Which statement is correct? Junior A. Experiment tracking — recording each training run's params, metrics and artifacts so results are comparable and reproducible B. Experiment tracking — applying DevOps practices to ML — automating the path from data to a monitored production model C. Experiment tracking — the serialized trained model plus metadata, promoted through registry stages to production D. Experiment tracking — the ability to recreate a model exactly by pinning data, code, environment and random seeds
7. What is Model lineage ? Mid A. running a candidate model against the incumbent on live data to decide promotion B. the recorded chain of data, code and config that produced a specific model version C. recording each training run's params, metrics and artifacts so results are comparable and reproducible D. the ability to recreate a model exactly by pinning data, code, environment and random seeds
8. Which term means: "the recorded chain of data, code and config that produced a specific model version"? Mid A. Model lineage B. MLOps C. Train-serve parity D. Experiment tracking
9. Which statement is correct? Mid A. Model lineage — reverting an inference endpoint to a previous model version when the new one underperforms B. Model lineage — the recorded chain of data, code and config that produced a specific model version C. Model lineage — the ability to recreate a model exactly by pinning data, code, environment and random seeds D. Model lineage — the serialized trained model plus metadata, promoted through registry stages to production
10. What is Reproducibility ? Mid A. reverting an inference endpoint to a previous model version when the new one underperforms B. recording each training run's params, metrics and artifacts so results are comparable and reproducible C. automating the train→validate→register→deploy steps as a repeatable, scheduled workflow D. the ability to recreate a model exactly by pinning data, code, environment and random seeds
11. Which term means: "the ability to recreate a model exactly by pinning data, code, environment and random seeds"? Mid A. Champion-challenger B. Reproducibility C. Experiment tracking D. MLOps
12. Which statement is correct? Mid A. Reproducibility — the recorded chain of data, code and config that produced a specific model version B. Reproducibility — ensuring the same transformation code runs in training and serving to avoid skew C. Reproducibility — automating the train→validate→register→deploy steps as a repeatable, scheduled workflow D. Reproducibility — the ability to recreate a model exactly by pinning data, code, environment and random seeds
13. What is Model artifact ? Mid A. the ability to recreate a model exactly by pinning data, code, environment and random seeds B. running a candidate model against the incumbent on live data to decide promotion C. recording each training run's params, metrics and artifacts so results are comparable and reproducible D. the serialized trained model plus metadata, promoted through registry stages to production
14. Which term means: "the serialized trained model plus metadata, promoted through registry stages to production"? Mid A. Model artifact B. Experiment tracking C. Rollback D. Reproducibility
15. Which statement is correct? Mid A. Model artifact — recording each training run's params, metrics and artifacts so results are comparable and reproducible B. Model artifact — the recorded chain of data, code and config that produced a specific model version C. Model artifact — the serialized trained model plus metadata, promoted through registry stages to production D. Model artifact — automating the train→validate→register→deploy steps as a repeatable, scheduled workflow
16. What is Pipeline orchestration ? Mid A. the serialized trained model plus metadata, promoted through registry stages to production B. automating the train→validate→register→deploy steps as a repeatable, scheduled workflow C. applying DevOps practices to ML — automating the path from data to a monitored production model D. reverting an inference endpoint to a previous model version when the new one underperforms
17. Which term means: "automating the train→validate→register→deploy steps as a repeatable, scheduled workflow"? Mid A. Model artifact B. MLOps C. Rollback D. Pipeline orchestration
18. Which statement is correct? Mid A. Pipeline orchestration — recording each training run's params, metrics and artifacts so results are comparable and reproducible B. Pipeline orchestration — the ability to recreate a model exactly by pinning data, code, environment and random seeds C. Pipeline orchestration — automating the train→validate→register→deploy steps as a repeatable, scheduled workflow D. Pipeline orchestration — reverting an inference endpoint to a previous model version when the new one underperforms
19. What is Champion-challenger ? Senior A. recording each training run's params, metrics and artifacts so results are comparable and reproducible B. the ability to recreate a model exactly by pinning data, code, environment and random seeds C. the recorded chain of data, code and config that produced a specific model version D. running a candidate model against the incumbent on live data to decide promotion
20. Which term means: "running a candidate model against the incumbent on live data to decide promotion"? Senior A. Reproducibility B. Experiment tracking C. Pipeline orchestration D. Champion-challenger
21. Which statement is correct? Senior A. Champion-challenger — recording each training run's params, metrics and artifacts so results are comparable and reproducible B. Champion-challenger — the serialized trained model plus metadata, promoted through registry stages to production C. Champion-challenger — the recorded chain of data, code and config that produced a specific model version D. Champion-challenger — running a candidate model against the incumbent on live data to decide promotion
22. What is Rollback ? Senior A. automating the train→validate→register→deploy steps as a repeatable, scheduled workflow B. reverting an inference endpoint to a previous model version when the new one underperforms C. the recorded chain of data, code and config that produced a specific model version D. recording each training run's params, metrics and artifacts so results are comparable and reproducible
23. Which term means: "reverting an inference endpoint to a previous model version when the new one underperforms"? Senior A. Pipeline orchestration B. Train-serve parity C. Rollback D. Experiment tracking
24. Which statement is correct? Senior A. Rollback — reverting an inference endpoint to a previous model version when the new one underperforms B. Rollback — running a candidate model against the incumbent on live data to decide promotion C. Rollback — the serialized trained model plus metadata, promoted through registry stages to production D. Rollback — recording each training run's params, metrics and artifacts so results are comparable and reproducible
25. What is Train-serve parity ? Senior A. applying DevOps practices to ML — automating the path from data to a monitored production model B. running a candidate model against the incumbent on live data to decide promotion C. recording each training run's params, metrics and artifacts so results are comparable and reproducible D. ensuring the same transformation code runs in training and serving to avoid skew
26. Which term means: "ensuring the same transformation code runs in training and serving to avoid skew"? Senior A. Rollback B. Train-serve parity C. Model lineage D. Champion-challenger
27. Which statement is correct? Senior A. Train-serve parity — automating the train→validate→register→deploy steps as a repeatable, scheduled workflow B. Train-serve parity — applying DevOps practices to ML — automating the path from data to a monitored production model C. Train-serve parity — running a candidate model against the incumbent on live data to decide promotion D. Train-serve parity — ensuring the same transformation code runs in training and serving to avoid skew