Feat/use jumpstart streamlined for new llm options

[aidanrussell]

Using Sagemaker Jumpstart containers and model weights stored in an AWS account in the eu-west-2 region, we redeploy the existing two small LLMs. This process then trivially allows for new LLMs to be added. Note that through jumpstart we also optimize the instance size and ensure the correct environment variables for the model, which can otherwise be difficult. It should be possible to also store these containers and model weights in our own account, I need to check how to do that, but I do wonder if it adds value since this way it never goes to the public internet in any case as it stays within the AWS region.