Professional Data EngineerOperationalizing machine learning modelsEasy
A data engineering team is developing a new machine learning model for fraud detection. The model will be deployed to predict fraudulent transactions in real-time. Due to the critical nature of fraud detection, the team needs to ensure that the model's predictions are highly reliable and that any degradation in performance is detected and addressed immediately. Which Google Cloud service is best suited for monitoring the model's performance in production?
- ACloud Monitoring
- BVertex AI Model Monitoring
- CDataflow
- DCloud Logging
Show answer & explanationAnswer & explanation
Correct answer: B. Vertex AI Model Monitoring
Vertex AI Model Monitoring is specifically designed to monitor machine learning models deployed on Vertex AI, providing features like drift detection and performance alerts. This directly addresses the need for immediate detection of performance degradation.
Why the other options are wrong
- A. Cloud Monitoring provides general infrastructure and application monitoring, but lacks ML-specific metrics and drift detection.
- C. Dataflow is a service for stream and batch data processing, not for monitoring deployed ML models.
- D. Cloud Logging is for collecting and viewing logs, not for model performance monitoring.
Vertex AI Model Monitoring
A Google Cloud service within Vertex AI that provides automated monitoring of deployed machine learning models to detect data drift, concept drift, and performance degradation.
- Detects data and concept drift.
- Monitors model performance metrics.
- Integrates with Vertex AI Endpoints.
Memory trick: Vertex AI watches your model's health, alerting you if it drifts or declines.