A2: How should I manage the data?

Ensure that data is accessible at scale and that systems integration is fully functional. This is particularly important for training or inference pipelines that require large datasets. 

Invest in proper systems integration, consider even streaming and queues to serve inference across systems in the organization.