So do you want Databricks to push data into Salesforce or do you want Salesforce to pull data?
If you want to push, you could write microbatches to salesforce (if possible of course).
For pull, you might wanna use the SQL api of databricks sql, OR use some kind of pub/sub tool like kafka etc where all changes are published, which can be fetched later on.

So perhaps some more info on 'Salesforce'.  Is it salesforce cloud or the database?