Deepgram
Deepgram builds voice-native AI foundation models and developer APIs for speech recognition, speech synthesis and low-latency voice interaction, used by 200,000+ developers and 1,300+ organizations.
Deepgram is a research-driven, foundational AI company that builds voice-native AI technology and developer tools for human-to-machine voice interaction. Its work spans speech recognition, speech-to-text, audio processing, speech synthesis, foundation models and low-latency inference. The company's stated aim is to make voice interaction instant, accurate and human-like.
Two products sit at the centre of its business. The first is a set of voice-native foundation models delivering speech recognition and voice interaction with, in the company's account, the lowest latency and the highest accuracy in voice AI, backed by enterprise-grade reliability at scale. The second is a developer voice AI platform, comprising tools and APIs for building conversational voice applications. Together these systems have processed more than 50,000 years of audio and over one trillion words.
Deepgram reports adoption by 200,000-plus developers and more than 1,300 organisations. Its customers include Twilio, Cloudflare, IBM and Sierra, spanning cloud communications and enterprise software. The company raised $130m in a Series C round and is valued at $1.3bn.
Roles centre on foundational AI research and the engineering of developer tools, with an emphasis on enterprise-grade reliability and scale.
