AutoML Overview

The AutoML subsystem automates the machine learning workflow from data preparation through model selection, hyperparameter tuning, and final model evaluation. It supports classification, regression, clustering, and time series tasks.

AutoML Pipeline

Data Preparation -> Algorithm Selection -> Training -> Hyperparameter Tuning -> Evaluation -> Model Saving

Job States

Status	Description
`pending`	Job created, not yet submitted
`queued`	Awaiting execution resources
`preparing`	Data loading and preprocessing
`training`	Model training in progress
`tuning`	Hyperparameter optimization
`evaluating`	Final model evaluation
`completed`	Pipeline finished successfully
`failed`	Pipeline encountered an error
`cancelled`	Job was cancelled

Creating an AutoML Job

POST /api/v1/automl/jobs
Content-Type: application/json
 
{
  "name": "customer-churn-automl",
  "task_type": "classification",
  "dataset_id": "dataset-123",
  "target_column": "churned",
  "feature_columns": ["tenure", "monthly_charges", "total_charges"],
  "algorithms": ["linear", "tree", "ensemble"],
  "tune_hyperparameters": true,
  "compute_type": "medium",
  "use_gpu": false
}

Source Files

File	Path
AutoML Orchestrator	`data-plane/ml-service/src/automl/automl_orchestrator.py`
Model Selection	`data-plane/ml-service/src/automl/model_selection.py`
Feature Engineering	`data-plane/ml-service/src/automl/feature_engineering.py`
NAS	`data-plane/ml-service/src/automl/nas.py`

Cost Management Model Selection