Hello @woop ,

We are running Kubeflow Pipelines 0.5 and are looking at using Feast as our feature store. Looking at the docs, it appears that BigQuery is used as the store for the feature data while Postgres is used to store feature metadata. Is there a hard dependency on BigQuery? Can we use any on-premise DB in its place to store the feature data? If so, can you please let us know how it can be done.