Thanks @shaun for this helpful and detailed response, and for everybody else (@woodpav, @david.wylie) for your valuable insights.
In my use case (long running data transfer process between an external API and a data warehouse) it definitely makes sense to use Anvil for the user interface/config and App Engine to run the process, however the data does not need to be sent back to Anvil (just status, which I can log somewhere else and access from my Anvil app) so I don’t think I need to use uplink in this instance, I can trigger the process via HTTP or Pub/Sub and keep it simple in Anvil.
Anyway, problem solved and Anvil looks like a good choice for this and future projects. Onwards!