Automating and orchestrating ML pipelines Questions
Practice questions for Automating and orchestrating ML pipelines topic in Google Professional Machine Learning Engineer. 44 questions covering this domain.
A team uses End-to-End AutoML for tabular modeling and wants to shrink the final ensemble model to improve serving latency and lower inference cost. W...
Which frameworks can be used to define ML pipelines for Vertex AI Pipelines?
Which statement correctly distinguishes a pipeline component from a pipeline task?
What is a pipeline run in Vertex AI Pipelines?
What best describes the structure of an ML pipeline in Vertex AI Pipelines?
A researcher wants to compare pipeline executions that used different hyperparameters and training steps inside one experiment context. What should th...
A team defines three pipeline tasks but does not connect them with input-output dependencies. How will Vertex AI Pipelines execute them by default?
A platform team wants one reviewed artifact from a pipeline definition that can be reused later for one-time runs or recurring scheduled runs. Which a...
A platform engineer needs to investigate which upstream artifacts and parameters produced a faulty model artifact from a pipeline run. Where is that l...
A team wants the same pipeline definition to run every night without manually starting each execution. Which Vertex AI capability should they use?
An engineer wants to include Vertex AI features such as AutoML directly in a pipeline definition without authoring every component from scratch. What ...
A pipeline must call a long-running Vertex AI hyperparameter tuning job inside a step and then use the best hyperparameters in a downstream training s...
Which approach is recommended to pass large datasets or model artifacts between Vertex AI Pipelines tasks?
Which Vertex AI Pipelines artifact captures the inputs, outputs, and execution history of a single pipeline run for later inspection?
Which Vertex AI Pipelines feature lets you reuse cached outputs from a previous task execution when inputs and code have not changed?
Which language is most commonly used to author Kubeflow Pipelines components that run on Vertex AI Pipelines?
A team wants a CI/CD trigger to compile a pipeline definition, upload it as a template, and start a pipeline run automatically when source code change...
A team wants to fan out a pipeline step to process partitions of a dataset in parallel within one pipeline definition. Which KFP construct supports th...
A team wants to start a pipeline whenever a new training data file lands in a Cloud Storage bucket. Which event-driven mechanism is appropriate?
Which approach is recommended to pass small parameter values between pipeline tasks in Vertex AI Pipelines?
Sign in to see all 44 questions
Create a free account to browse all questions — completely free during our launch phase.