Hi, i know how to visualise and track results for single flows (that works great, html is really flexible !). But what about lets say we did run a machine learning flow 10 times with different params, tracked the accuracy, and now want to see which one of them performed best. Whats the easiest/best way to do that in metaflow, or does it make more sense to use an external tool like mlflow here ?

You can do it playing around with metaflow API, but probably better mlflow and similar.