AI Platform

Google Cloud AI Platform (now part of Vertex AI) is a managed machine learning platform that enables data scientists and ML engineers to build, train, deploy, and manage ML models at scale. It provides a unified environment for the entire ML workflow.

Key Features of AI Platform:

1. Training - Run distributed training jobs using custom containers or built-in algorithms.

2. Prediction - Deploy trained models for online and batch predictions.

3. Notebooks - Managed JupyterLab instances for experimentation.

4. Pipelines - Orchestrate end-to-end ML workflows using Kubeflow Pipelines.

5. Feature Store - Centralized repository for storing and serving ML features.

The below example shows how to submit a training job to Google Cloud AI Platform using the Python SDK.

It gives the following output,

Job created: my_training_job_001
State: QUEUED

AI Platform Training Job States:

1. QUEUED - Job has been submitted and is waiting to be scheduled.

2. PREPARING - Resources are being allocated for the job.

3. RUNNING - The training job is actively running.

4. SUCCEEDED - The job completed successfully.

5. FAILED - The job encountered an error and stopped.

AI Platform has now evolved into Vertex AI, which provides a unified platform combining AutoML and custom training with additional MLOps capabilities including model monitoring, experiment tracking, and managed pipelines.