How do you address the limitation of LLM-as-a-judge when evaluating nuanced clinical decisions at Amigo?
