Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
An enterprise is deploying trained machine learning models on Google Cloud to power two distinct business applications:
Which Vertex AI serving methods should the organization use for Application 1 and Application 2?
Vertex AI provides two primary mechanisms for serving model predictions: online prediction endpoints and batch prediction jobs. Each is optimized for distinct operational and latency requirements within the machine learning lifecycle.
Pairing persistent online prediction endpoints with interactive user applications while utilizing batch prediction for asynchronous bulk workloads follows Google Cloud's architectural best practices, optimizing both performance and operational cost.
AutoML and custom training are model development and training methodologies in Vertex AI, not model deployment or prediction serving mechanisms.
The question asked how to train a model without writing machine learning code versus developing a complex model architecture with custom dependencies and frameworks.
Vertex AI Feature Store and Vertex AI Model Registry are components for data management and model governance, not prediction serving engines.
The scenario required centralizing tabular features for real-time retrieval and maintaining an enterprise catalog of versioned models ready for deployment.
This option reverses the intended design patterns of batch prediction and online prediction endpoints.
Application 1 was an asynchronous background task processed overnight and Application 2 was an interactive real-time API.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated