Key Concept Anchor Points

This lesson provides a high-level review of the key concepts for each domain of the exam. Use this as a final checklist to identify any areas where you may need to do some last-minute review.

1. Architecting Low-Code AI Solutions

• BigQuery ML vs. AutoML vs. Custom: Know when to use each of these tools.
• Pre-trained APIs: Understand the capabilities of the Vision, Natural Language, Translation, and Speech APIs.

2. Collaborating on Data and Model Management

• Dataflow: Know how to use Dataflow for batch and streaming data processing.
• Vertex AI Feature Store: Understand the difference between the online and offline stores and how to use the Feature Store to prevent training-serving skew.

3. Scaling Prototypes into ML Models

• Vertex AI Workbench: Know the difference between managed and user-managed notebooks.
• Vertex AI Experiments: Understand how to use Vertex AI Experiments to track and compare your experiments.

4. Serving and Scaling Models

• Batch vs. Online Prediction: Know when to use each of these serving patterns.
• Vertex AI Prediction: Understand how to deploy models to endpoints, manage versions, and optimize for latency.

5. Automating and Orchestrating ML Pipelines

• Vertex AI Pipelines (KFP): Know how to define pipeline components and how to use the KFP SDK to create and run pipelines.
• TFX: Understand the key components of a TFX pipeline and when to use TFX instead of KFP.

6. Monitoring ML Solutions

• Vertex AI Model Monitoring: Know how to set up monitoring for your models and how to interpret the results.
• Training-Serving Skew vs. Prediction Drift: Understand the difference between these two types of drift and how to detect them.

Knowledge Check