Is there a way to split a trpc api across multiple lambdas ?

How do I go about splitting my TRPC api across multiple lambdas, such that each lambda would load the minimum code it needs to run, (ideally the frontend should not be affected )