Google Professional Machine Learning Engineer Question 344
Single answerGoogle Cloud PlatformYou are a machine learning engineer tasked with deploying a trained model for real-time inference on Google Cloud. The model has been packaged in a Docker container. The solution must allow for horizontal scaling and seamless integration with other services. Which Google Cloud service should you use to serve the containerized model?
- A
Cloud Run
- B
Google Kubernetes Engine (GKE)
- C
AI Platform Prediction
- D
App Engine
Show answer and explanation
Correct answer: A
Explanation
Cloud Run is the best choice for serving containerized models because it is fully managed, supports containerized deployments, and provides automatic scaling. It simplifies operational complexity compared to GKE and outperforms AI Platform Prediction and App Engine for this specific use case.
- A. Correct.
Cloud Run is a fully managed compute platform that supports deploying containerized applications. It is ideal for serving containerized models due to its automatic scaling, support for containerized workloads, and simple integration with other Google Cloud services.
- B. Incorrect.
Google Kubernetes Engine (GKE) is a powerful container orchestration platform, but it requires manual setup and management of Kubernetes clusters, which adds operational overhead compared to Cloud Run for simple containerized serving tasks.
- C. Incorrect.
AI Platform Prediction is designed for serving machine learning models, but it does not directly support serving arbitrary Docker containers. It is more suitable for TensorFlow or Scikit-learn models rather than custom containerized solutions.
- D. Incorrect.
App Engine is primarily designed for serving web applications and does not natively support containerized workloads as efficiently as Cloud Run does.