Deepgram

Deepgram was founded in 2015 by Scott Stephenson, Adam Sypniewski, and Noah Shutty—all former physicists at the University of Michigan. Fascinated with the idea of life-streaming, Scott and Noah built wearable recording devices, on which they recorded hundreds of hours of their voice notes, conversations, and chance interactions as they went about their days. This presented a problem: how to sift through all that audio data. Translating his research work on “listening” for signatures of dark matter interactions using end-to-end deep learning, Scott built a deep learning model capable of searching through voice data. The three founders left academia to move to the Bay Area and start Deepgram.

Deepgram makes voice data useful to people and usable by computers using end-to-end deep learning models. Speech is the dark matter of enterprise data and is a largely untapped source of insights and business value. We get to help our customers discover that value and advance the state of the art in deep learning and language models.