# Discerning sentence boundaries in speech-to-text of long paragraphs

This seems like it'd be a common issue in speech to text systems (but I'm not very familiar with the area):

If you're speaking long paragraphs, it's hard to discern sentence boundaries (e.g. if I pause in the middle of a sentence)

Right now I'm using pauses for sentence boundaries. That causes outputs like this:

My friend worked. On a cool hackathon project


What are some techniques I could use to perhaps post-process this STT output to convert it to the following?

My friend worked worked on a cool hackathon project.