Model Optimization Quickstart

Model Optimization supports iterative tuning from model registration through experiment validation and production promotion. Teams get the best outcomes when they use consistent naming, documented parameter intent, and staged rollout discipline.

What to Click and Do Next

1. In the sidebar, click Apps > Model Optimization.
2. Click Create to register a model.
3. In Register a new model, enter Name, Product Family, Rule Type, and Description, then click Save.
4. Open the Parameters tab.
5. Click New Parameter and configure error-type behavior.
6. Save parameter updates.
7. Open Experiments and click New Experiment.
8. Create and validate experiment set.
9. Open Productions and click New Production.
10. Configure production set and save.
11. Use Settings tab for metadata updates or controlled retirement.

Validation Checklist

1. Model appears with expected metadata.
2. Parameters are saved and visible.
3. Experiment and production sets open and run as expected.

Common Issues

• Module or actions hidden due to feature gating or insufficient role.
• Poor tuning outcomes from unclear parameter baselines.