Goal: 

Goal based statement, with context. Can take the form of “As a <User/Actor>, I Want <to Achieve Some Goal>, so that <Some Reason/Context>”.

• The quickest way to distributed training via LLS is to use the Kubeflow trainer which will kickoff a torch run job using the instructlab training arguments.
• build an out of tree provider: https://github.com/opendatahub-io/llama-stack-provider-kft which enables users to use the llama-stack-client SDK and CLI to kick off a post_training run using KFTO compatible with InstructLab SDG

Acceptance Criteria:

A list of verification conditions, successful functional tests, or expected outcomes in order to declare this story/task successfully completed.

• provider can be used inside of upstream LLS
• provider is packaged and installable
• the LLS SDK can successfully orchestrate a post_training run using this provider.