These are important tasks in MLOPS

Based on this we have the following best practices

Train model

Data and ML pipelines automation

Experiment management and metrics tracking

Model registry

Automate CI/CD