It is probably best to use a different mental model when trying to understand forge. It is not a typical “Serverless” or FaaS platform.
Although forge runs on AWS and uses familiar terminology like “handler”, there is little similarity between the sandboxed environment of forge and AWS lambda. You should not approach writing forge functions as if you’re writing AWS lambda functions.
Even though @denizoguz is right, and you can invoke an http call to an AWS lambda function from within a forge function handler, i would not recommend doing so if the sole purpose is to cling on to a familiar mental model for writing serverless functions.
I would say that it is best to spend time learning the intricacies of the forge platform, preferably without the burden of previously learned patterns, and create native forge applications.