Founding Researcher

Miso Labs•San Francisco, CA

11h•Onsite

About The Position

Miso Labs is hiring founding researchers to help develop full-duplex speech-to-speech models and advance its goal of passing the voice Turing test. The work involves exploring new model architectures and scaling them rapidly, spanning the model-training stack from small-scale architecture experiments to large pre-training runs on up to 1,000 H100s. Researchers are encouraged to publish work outside the company’s core intellectual property, and the company will sponsor travel to leading AI/ML conferences. This is an in-person role in San Francisco, and employees are expected to relocate to the San Francisco area; moving costs are covered. The company emphasizes giving team members credit for their work.

Requirements

Excitement about working on all aspects of model training and development.
Comfort with both research and hands-on engineering.
Ability to own long-term research outcomes end-to-end.
Willingness to learn technical skills quickly and in a self-directed manner.
Prior experience with both PyTorch and CUDA.

Nice To Haves

A bachelor’s degree in a technical field such as Computer Science, Physics, or Mathematics, or equivalent experience.

Responsibilities

Explore and develop architectural ideas for full-duplex speech-to-speech models.
Work across the model-training stack, from small-scale architecture experiments to large pre-training runs.
Own long-term research outcomes end-to-end.
Publish research outside the company’s core intellectual property where appropriate.