I am currently building an production grade application, where we make multiple llm calls inside the custom llm. We are working on to improve the speed. any advice would be good. There are the options I find valuable - Using BackgroundTask ( sry, we are using FastAPI ) - Semantic Cache with redis - Map Reduce ( not sure ~ ) - MCP Servers ( not sure ~ ) Any suggestions would be more helpful. What do you think?

Hey! I couldn't find enough information to answer your question, so I'm escalating this to a human agent who can help you better. They'll get back to you soon. Meanwhile, here's a bit about Custom LLMs in Vapi: A Custom LLM in Vapi is a conversational assistant that can call external functions, trigger processes, and handle special logic, all while chatting with your users [source](https://discord.com/channels/1211482211119796234/1328028159949537341). 😊

Okay ducky!

checking if this is resolved/solved for you?