Developing ML solutionsFull notesSummaryRingkasanStoriesPracticeFundamental Concepts of MLOps10 exam-style questions on this lesson.Question 1 of 10What is MLOps?AA type of neural network for operations dataBA way to label training data fasterCA specific AWS service for hosting modelsDApplying DevOps practices to the whole ML lifecycleCheck answerQuestion 2 of 10Why does ML need its own operations practices beyond standard DevOps?AModels degrade as the data changes over timeBML models can't be stored in version controlCML code doesn't need automated testingDML code never changes after the first releaseCheck answerQuestion 3 of 10Which is unique to CI/CD for ML compared to traditional software CI/CD?AContinuous delivery, which releases new versionsBContinuous training, which retrains models automaticallyCVersion control, which tracks code changesDContinuous integration, which merges code oftenCheck answerQuestion 4 of 10Why should a team version its data and models, not only its code?ABecause AWS requires it for billingBTo make training fasterCFor reproducibility and the ability to roll backDTo save storage costsCheck answerQuestion 5 of 10. Select two.Which TWO are goals of MLOps? (Select TWO.)AEncourage collaboration between data scientists, engineers, and IT operationsBDeploy without any reviewCRemove the need for data scientistsDSpeed up the lifecycle through automationEAvoid monitoring to save costsCheck answer0 of 2 selectedQuestion 6 of 10What does model governance in MLOps include?ARemoving logs to protect model performanceBOnly choosing which framework the model usesCLetting any user deploy any model version directlyDDocumentation, review and approval, data protection, and complianceCheck answerQuestion 7 of 10In a production MLOps setup, how are training and deployment usually organized?ASeparate build and deployment pipelines, with new data triggering buildsBDeployment happens first, then training in productionCOne manual script trains and deploys each releaseDTraining is done only once, before the first launchCheck answerQuestion 8 of 10A team wants to orchestrate its end-to-end ML workflow on AWS. Which service fits?ASageMaker Feature StoreBSageMaker PipelinesCSageMaker Model MonitorDSageMaker Ground TruthCheck answerQuestion 9 of 10. Select two.Which TWO AWS options can prepare data in an MLOps workflow? (Select TWO.)AAmazon PollyBAWS ArtifactCSageMaker Processing APIDSageMaker Model RegistryESageMaker Data WranglerCheck answer0 of 2 selectedQuestion 10 of 10What does continuous monitoring track in MLOps?AOnly the number of active usersBOnly the final training lossCData, models, and business metricsDOnly the uptime of the serversCheck answerModel Deployment8 exam-style questions on this lesson.Generative AI Application Lifecycle6 exam-style questions on this lesson.