Hello, is Apache Flink planning to add support for LLM's? We can run LLama using something akin to the dalai package with enough memory, will an integration be possible on streaming data in the future? I'm thinking of transcribing a speech to text to LLama

This doesn't sound like something Flink has to explicitly support. You can use use async i/o to reach out to external apis, or use a map or process function to use any library of your choosing to do data enrichment.