Has anyone built an analytics stack with Tuva where they use Medplum for their FHIR server? Would love to understand how you set it up / any resources you referenced.

I have not done this personally, so caveat but Tuva supports NDJSON (https://github.com/tuva-health/FHIR_inferno) and so does Medplum https://www.medplum.com/docs/api/fhir/operations/bulk-fhir So in theory, you could bulk export and then input into the Tuva engine.