As per our high level customer journey map, a reasonable upgrade path from single-node hosted notebooks to a distributed pipeline on RHOAI could be bringing our deep dive notebooks into an Elyra pipeline. Several questions need to be investigated:

• Data access: What would it take to make our notebooks be able to read from data made available in RHOAI? 
  • Note: this might be almost a noop if it's as simple as reading from a path on a PVC
  • Might have to add supporting setup / documentation for how to discovery file locations on RHOAI and associate with RHOA data import documentation
• What is the best mechanism for creating a pipeline of our notebooks on RHOAI?
  • Context: the apparently planned replacement of Elyra with KFP. What would the lift be? Would the interface remain?
• What might it take to adapt our notebooks to use Llama Stack clients? Which operations would need it and while would not? 
  • What about docling-serve for the distributed case?

Acceptance criteria:

• Google Doc with discoveries and possible solutions along with their tradeoffs