Speech and Audio Researcher - Remote

Overview

Homebrew is looking for an experienced Researcher to contribute to our speech and audio efforts across TTS, ASR, and LLMs.

In Short

• Work at the intersection of speech, audio & robotics
• Build speech models, pipelines and data using TensorFlow, PyTorch, Torchtune, etc.
• Stay up-to-date with the latest advancements in speech research, including flow matching, across fusion techniques, etc
• Collaborate with cross-functional teams to rapidly productionize research
• Publish data sets, models, and papers

Requirements

• Strong programming skills in Python, C++, or Java, with experience working with deep learning frameworks and libraries
• In-depth knowledge of AI fundamentals
• Familiarity with some of the following: TTS/STT, ASR, CSR, encodecs, VAD, noise cancellation, duplex systems, sentiment/emotion/noise detection, etc.
• Up-to-date knowledge of speech & audio techniques, e.g. flow matching, diffusion, etc
• Bonus: relevant publications or production-level speech work

Benefits

• Remote work with a documentation-first, open-source, minimal meeting culture
• 14 days leave (and unlimited sick days)
• Perks: subscriptions, equipment budget, speaking coaches, and more