AI Project Life Cycle, MLOps, and DataOps CAIPM Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Which DataOps principle emphasizes the use of automated processes to reduce manual errors and speed up data pipeline delivery?

   • AAutomation
   • BVersion control
   • CMonitoring
   • DCollaboration
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A telecom company is implementing MLOps for its churn prediction model. The team is struggling with reproducibility because data scientists use different environments and dependencies. Which best practice should they adopt to improve reproducibility?

   • AUse containerization (e.g., Docker) for training and serving environments.
   • BStandardize on a single cloud provider for all ML workloads.
   • CPin the model's hyperparameters in a configuration file.
   • DRequire data scientists to document their environment manually in a wiki.
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   A media company's recommendation model relies on a real-time data pipeline that ingests user clickstream data. The data engineering team has set up automated data quality checks, but the ML team complains that the model's performance degrades during peak traffic hours. Investigation shows that the pipeline drops some events during high load, but the data quality checks only run on a sample. Which change would best address the issue?

   • AIncrease the model's retraining frequency during peak hours
   • BReduce the volume of data ingested during peak hours
   • CRun data quality checks on the full data stream, not just a sample
   • DHave the ML team manually review the data during peak hours
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   An e-commerce company has a team of data scientists who develop models locally and then hand off code to a separate team for deployment. The handoff often fails because the data scientists use different library versions and the deployment team cannot reproduce the training environment. Which tool should the company adopt to ensure environment consistency?

   • AA monitoring tool (e.g., Prometheus)
   • BA version control system (e.g., Git)
   • CContainerization (e.g., Docker)
   • DA data orchestration tool (e.g., Apache Airflow)
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A government agency is building a model to predict traffic congestion. The data engineering team uses an orchestration tool to run data pipelines that pull data from multiple sources. The team wants to ensure that if a pipeline step fails, the downstream steps are not executed, and the issue is automatically reported. Which DataOps feature should they configure?

   • AParallel execution of all pipeline steps
   • BManual rerun of the entire pipeline
   • CData quality checks only at the end of the pipeline
   • DDependency-based scheduling with failure alerts
   Reveal Solution Discussion Select an answer first