Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
Comments
Loading comments...
Ultimate access to all questions.
Your company operates a global e-commerce platform that utilizes TensorFlow-based deep learning models for processing Analytics-360 data to personalize user experiences. Initially, these models significantly improved user engagement and sales. However, over the past six months, the models' performance has degraded due to evolving user behavior and market trends. After manually retraining the models with updated data, performance improved, but the manual process is not scalable. Your team is now tasked with automating the model retraining and deployment process within Google Cloud, ensuring minimal downtime and cost efficiency. The solution must support continuous integration and delivery (CI/CD) for machine learning models, provide version control for models, and automate the monitoring of model performance. Which of the following solutions best meets these requirements? (Choose two correct options if option E is available, otherwise choose one.)
A
Deploying the models on Google Kubernetes Engine (GKE) and using TensorFlow Extended (TFX) for automating the ML pipeline, including data validation, model training, and deployment.
B
Utilizing Google Compute Engine clusters with Kubeflow for orchestrating the ML workflows, but without integrated data validation or model monitoring features.
C
Leveraging Google AI Platform with TFX to manage the end-to-end ML lifecycle, including automated data validation, model training, evaluation, and deployment, with built-in performance monitoring.
D
Implementing a custom solution on Google Cloud Functions for model retraining and deployment, requiring significant manual intervention for data validation and model monitoring.
E
Combining Google AI Platform for managed services with Kubeflow on GKE for flexibility in orchestrating complex ML workflows, including data validation, model training, and performance monitoring.