AI Researcher (Voice)

TL;DR

AI Researcher (Voice): Leading research and development of generative audio and speech models for real-time human simulation with an accent on flow matching, diffusion architectures, and audio-to-expression synthesis. Focus on productionizing cutting-edge research, innovating streaming speech models, and driving advancements in multimodal AI.

Location: Preferably hybrid in San Francisco with relocation provided, but open to global remote candidates.

Compensation: $160,000–$250,000

Company

hirify.global is a research lab building AI Humans—a new interface for real-time, face-to-face conversations between people and machines.

What you will do

• Lead research efforts on generative video and audio models, including text-to-speech and speech-to-speech.
• Collaborate with the Applied ML team to bridge the gap between research prototypes and production systems.
• Stay current with latest AI advancements and publish original research in top-tier venues.
• Innovate on representation learning architectures in audio and image domains.
• Prototype and test novel ideas for expressive, lifelike avatars.

Requirements

• Proven experience with flow matching, diffusion models, and auto-regressive networks in the audio domain.
• Demonstrated ability training large-scale deep learning models.
• Strong programming skills with high fluency in PyTorch.
• Track record of original research with publications in venues like CVPR, NeurIPS, or BMVC.
• Experience building streaming text-to-speech or speech-to-speech models.
• Ability to thrive in a fast-paced startup environment and take ownership of complex research paths.

Nice to have

• PhD or equivalent experience.
• Knowledge of 3D graphics and Gaussian splatting.
• Experience leading research teams.
• Strong background in software development best practices.

Culture & Benefits

• Flexible work schedule and unlimited PTO.
• Competitive healthcare coverage.
• Gear stipends for work setup.
• Supportive, diverse team culture focused on creating, not just fitting in.