What You Should Know:
– Voice AI research company Hume AI announced a significant leadership transition and strategic expansion of its research infrastructure business.
– Andrew Ettinger has been appointed as the new Chief Executive Officer, taking over as co-founder Alan Cowen moves to Google following a technology licensing agreement between the two companies.
Leadership and Strategic Shift
Andrew Ettinger, formerly Chief Revenue Officer at Appen, brings 15 years of experience in scaling AI data and infrastructure teams. His appointment signals Hume AI’s evolution from a model-focused startup into a primary research infrastructure partner for leading AI labs and enterprises.
- Financial Momentum: Hume AI is on track to generate over $100M in revenue this year, driven by demand for its voice data, annotations, and post-training workflows.
- Google Partnership: Google has non-exclusively licensed certain Hume technologies to bolster Gemini’s voice and emotion capabilities.
- Talent Acquisition: Alongside Cowen, roughly seven senior engineers from Hume AI are joining Google DeepMind.
“Voice in AI is evolving from a feature to the primary interface for the next generation of applications and devices,” said Ettinger. “Understanding emotion will be essential to unlocking AI’s full potential and that will require ongoing systems that incorporate human-in-the-loop feedback. That’s where Hume’s data, annotation, and reinforcement-learning infrastructure is setting the pace for the industry.”
Core Research Infrastructure
The company is doubling down on its “voice gym” infrastructure, providing the tools necessary for other labs to build emotionally intelligent voice models.
- Voice Evaluation Stack: Hume offers proprietary datasets and evaluation software to help hyperscalers and AI labs tune voice assistants.
- Reinforcement Learning: Its “voice gym” uses Reinforcement Learning from Human Expression (RLHE) to train models to be more realistic and empathic.
- Next-Gen Models: Despite the talent move to Google, Hume remains independent and plans to release its next generation of text-to-speech (TTS) and speech-to-speech (STS) models in the coming weeks.
