Create plan for hosted inference for Red Hat associates

[owtaylor]

Many Red Hat developers have Intel laptops with a weak iGPU that can't run our models effectively. Allowing them to test Granite.Code would:

• Increase the feedback we get
• Allow them to learn about AI and AI coding assistance
• Provide coding assistance (especially now that we have better autocomplete)

We should come up with a plan for running an inference server internally - probably on top of existing infrastructure - and a way to sign up and download config files.

(Sorry - not something we can offer externally - though it should be replicable given the resources.)