rhoai-automl

name	rhoai-automl
metadata	{"author":"rhoai3-demo","version":"1.0.0","platform-family":"rhoai","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"RHOAI Platform"}
description	Use when documenting, reviewing, or operating Red Hat OpenShift AI AutoML workflows from the official Working with AutoML guide: Technology Preview posture, CSV training data in S3-compatible storage, AutoML optimization runs, binary classification, multiclass classification, regression, time series forecasting, externally imported AutoML pipeline naming, leaderboard evaluation, feature importance, confusion matrices, model registry handoff, saved prediction notebooks, AutoGluon ServingRuntime deployment, task metrics, and AutoML configuration parameters. Do NOT use for generic AI Pipelines server administration (use rhoai-ai-pipelines), S3 object operations outside AutoML (use rhoai-s3-object-storage-data), project/workbench lifecycle (use rhoai-project-workflows), IDE workflows outside saved AutoML notebooks (use rhoai-data-science-ide-workflows), model registry administration (use rhoai-model-registry), general model-serving platform configuration (use rhoai-model-serving-platform), or live cluster changes without the OpenShift safety guard.

RHOAI AutoML

Use this skill for Red Hat OpenShift AI AutoML user workflows on the active product baseline in docs/PLATFORM_BASELINE.md.

Source Grounding

Read references/source-capture.md before using product workflow details. Official Red Hat documentation is product authority. This skill adapts the official Working with AutoML guide to this repo's demo workflow and governance review model.

Scope

This skill covers:

• AutoML Technology Preview support posture
• AutoML concept model and workflow
• supported task types: binary classification, multiclass classification, regression, and time series forecasting
• CSV training data requirements and size limits
• S3-compatible object storage and AI Pipelines server prerequisites
• AutoML optimization run creation and immutable run configuration
• externally imported AutoML pipeline naming rules
• leaderboard evaluation, model details, feature importance, and confusion matrix boundaries
• model registry handoff from the AutoML leaderboard
• saved notebook prediction workflow from a workbench
• AutoGluon ServingRuntime prerequisite and deployment handoff
• time series runtime environment variables for non-default column names
• optimized metrics and configuration parameters by task type

Use other skills for adjacent work:

• rhoai-ai-pipelines for pipeline server configuration, pipeline run lifecycle, imported pipeline management, caching, logs, and DSPA troubleshooting
• rhoai-s3-object-storage-data for object storage operations from workbenches
• rhoai-project-workflows for projects, workbenches, connections, project access, and cluster storage
• rhoai-data-science-ide-workflows for running saved notebooks and Git or Python package workflows inside a workbench
• rhoai-model-registry for administrator provisioning of model registries
• rhoai-model-registry-workflows for registered model/version metadata, deployment handoff, archive, and restore
• rhoai-model-serving-platform for ServingRuntime review and serving platform configuration
• rhoai-model-management-monitoring for day-2 deployed model operations and monitoring
• rhoai-api-tiers for API support posture, including KServe serving.kserve.io/v1alpha1 resources

Demo Policy

For this repo:

• Label AutoML as Technology Preview in READMEs, runbooks, presentations, and demo scripts. Do not present it as production SLA-backed.
• Treat AutoML as a rapid model exploration and comparison workflow for tabular CSV data, not as a replacement for governed pipeline engineering.
• Require a data science project with an AI Pipelines server before any AutoML run is promised.
• Store training data in S3-compatible object storage and keep object storage credentials in project-scoped connections or Secrets.
• Keep training CSV files UTF-8 encoded, comma-delimited, and header-bearing.
• Respect the documented data size limits: 32 MiB for dashboard upload and 100 MB when loaded from S3-compatible storage.
• Do not claim custom algorithm or hyperparameter selection; AutoML chooses algorithms, hyperparameters, train/test split, and optimization metric.
• Treat an optimization run as immutable after creation. Use AI Pipelines run workflows to stop, archive, or delete the underlying pipeline run.
• Register selected models to a model registry before deployment when the demo needs a governed handoff.
• Use saved notebooks for transparent prediction exploration in a workbench.
• Deploy AutoML models only after the AutoGluon serving runtime is available and its Red Hat registry image and KServe schema are verified.
• For time series forecasting, set runtime environment variables when the trained data used non-default series ID or timestamp column names.

Workflow

1. Confirm the active baseline in docs/PLATFORM_BASELINE.md.
2. Read references/source-capture.md and references/official-doc-extraction.md.
3. Confirm Technology Preview posture is acceptable for the requested demo surface.
4. Decide whether the task is:
   • AutoML concept or README authoring
   • prerequisite review for project, AI Pipelines, S3 data, and cluster capacity
   • optimization run design
   • external AutoML pipeline import review
   • leaderboard evaluation and model selection
   • model registry or saved notebook handoff
   • AutoGluon serving runtime or deployment review
   • metric and parameter interpretation
5. Use examples/automl-patterns.md for focused review patterns.
6. For live cluster work, follow the OpenShift safety guard in AGENTS.md.
7. Validate with references/validation-checklist.md.

References

• references/source-capture.md
• references/official-doc-extraction.md
• references/validation-checklist.md
• examples/automl-patterns.md