What is a Research Scientist at Deepgram?
A Research Scientist at Deepgram works at the absolute frontier of speech artificial intelligence and deep learning. In this role, you are not merely applying existing models; you are designing, training, and optimizing the next generation of end-to-end automatic speech recognition (ASR), text-to-speech (TTS), and natural language understanding (NLU) systems. Your work directly dictates the speed, accuracy, and capabilities of the Deepgram platform, which processes millions of hours of audio for global enterprises.
The impact of this role is immediate and highly visible. By developing novel neural network architectures and innovative training methodologies, you directly reduce latency and word error rates (WER) across Deepgram's production APIs. You will work with massive, diverse datasets to build models that perform robustly across various accents, noisy environments, and specialized domain vocabularies, ensuring that state-of-the-art research translates seamlessly into real-world utility.
This position is ideal for researchers who thrive on the synthesis of theoretical innovation and practical engineering. At Deepgram, you will bypass the slow research cycles of traditional academia to rapidly deploy your models into a high-throughput production pipeline. It is a highly collaborative, fast-paced environment where your discoveries will shape how humans interact with machines through voice.



