Google Associate Cloud EngineerDeploying and implementing a cloud solutionHard

A client needs to deploy a custom machine learning model for real-time inference. The model is packaged as a Docker container and requires GPU acceleration for optimal performance. They also need to monitor the model's performance and manage different model versions. Which Google Cloud service should be used?

  1. AGoogle Kubernetes Engine (GKE) with GPU nodes
  2. BCloud Run with GPU acceleration enabled
  3. CCompute Engine instances with GPUs and custom deployment scripts
  4. DVertex AI Prediction
Show answer & explanation

Correct answer: D. Vertex AI Prediction

Vertex AI Prediction is a fully managed service designed for deploying and serving machine learning models, including those packaged as Docker containers and requiring GPU acceleration. It provides built-in features for model monitoring, version management, and auto-scaling, significantly reducing operational overhead compared to self-managing infrastructure.

Why the other options are wrong

  • A. GKE can host containers with GPUs, but Vertex AI Prediction offers a higher-level, ML-specific managed service with integrated features like monitoring and versioning.
  • B. Cloud Run currently does not support GPU acceleration for inference.
  • C. Compute Engine requires significant manual effort for deployment, scaling, monitoring, and version management of ML models with GPUs.

Vertex AI Prediction

A fully managed service within Google Cloud's Vertex AI platform for deploying and serving machine learning models for online (real-time) predictions.

  • Supports custom containers for model deployment.
  • Can leverage GPU acceleration for inference.
  • Provides model monitoring and version management.
  • Offers auto-scaling based on traffic patterns.
  • Integrates with other Vertex AI services for MLOps.

Memory trick: Vertex AI Prediction: It's the 'peak' for putting your machine learning models to work, smart and fast.

More Deploying and implementing a cloud solution questions