An app should use a model hosted by Databricks, with no serving setup. What fits?
Why A is the answer
Foundation Model APIs
Foundation Model APIs give you hosted models to call, with no serving infrastructure to set up.
In detail
- Foundation Model APIs give you hosted models.
- You call them without setting up serving yourself.
- Check that the model you need is offered there.
Use case
An illustrative scenario, not a real incident.
A team needs a hosted LLM quickly. Foundation Model APIs give an endpoint to call, so nobody has to set up GPUs or serving.
See it in different ways
sequenceDiagram participant C as pyfunc chain participant M as MLflow participant U as Unity Catalog participant S as Model Serving C->>M: Log the model M->>U: Register a version U->>S: Deploy to an endpoint
flowchart TD
Q{"What do you need to manage?"}
Q -->|"Nothing"| A0["Foundation Model APIs"]
Q -->|"Your own custom model"| A1["Model Serving"]Key point: Foundation Model APIs give hosted models with no serving setup.
Why the other options are wrong
- B. A self-managed GPU cluster
- A GPU cluster is exactly the setup you want to avoid.
- C. A custom container on a VM
- A custom container adds work to run.
- D. A model loaded on the driver node
- A driver node is not a serving layer.