[META] Generic ML inference processor for search pipeline

[ylwu-amzn]

Related issue #2414

We have RAG and rerank search pipeline today. Some user want to leverage Gen AI in other use case, for example, they can leverage lang-ident model to check what language user is using, then translate the search response to that language. We will build generic ML inference search processor for search request and response, so customer can use any model in their search pipeline.