A simple query parser built to decrease workloads for LLMs in the performance of a chatbot.

Platform: DiFy

Problem: LLM overused considerably slowed down the chatbot's responses (+~20 secs per query), hence the need to integrate code as much as possible

Scenario: When a followup question is asked, instantly fuse the queries to reduce computation need for LLM