An ML pipeline runs with an administrator identity because multiple steps need different resources. Which approach best fits?

1. Bind runs to versioned input data and capture configuration.
2. Design idempotent steps and durable completion tracking.
3. Use appropriately scoped workload identities and permissions for the stages.
4. Define controlled retraining triggers and validation steps.