I became aware of LlamaIndex due to the possibility of providing langchain conversational memory a hypothetical infinite memory. I'm having a hard time using this in my agent which was originally an LLMChain. Does anyone have a few more examples of langchain + llamaindex?