2 comments

[ 2.7 ms ] story [ 12.9 ms ] thread
Interesting approach with the cascade. How do you decide when to escalate from fuzzy matching to LLM?