I'm not sure if I'm stating something obvious here or whether this is thourougly documented. I have tried to search on this for a long time but to no avail.
We are running our machine learning models as data pipelines with data ingestion > training > prediction.
However, instead of creating a new flow/deployment back test for each of these models, I would prefer to create only one flow where each of the models can be imported using paramatization. How would you approach this task?