Physical AI Researcher

Feather Robotics; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Feather Robotics in San Francisco seeks a Physical AI Researcher to lead applied research advancing reliability and data efficiency for manipulation in general-purpose robots.

You will own the research stack, collaborate with engineers to integrate perception, control, and cognition, and push Physical AI toward mainstream commercial deployment through rigorous experiments and publications.

Qualifications

• MS or PhD in EECS, Robotics, CS, AI, or related field, or equivalent industry experience
• Strong background in machine learning, with applied work in robotics or embodied AI
• Experience with one or more of: Vision-Language Models (VLAs), imitation learning, diffusion policies, reinforcement learning, navigation
• Proficiency in Python and modern ML frameworks (e.g., PyTorch, TensorFlow)
• Track record of impactful projects, whether in academia, open-source, or industry
• Publications in top robotics conferences (ICRA, CoRL, RSS, IROS, etc.)

Responsibilities

• Develop, train, and deploy machine learning models for general-purpose robots
• Implement and optimize algorithms in areas such as imitation learning, RL, diffusion policies, VLA, and navigation
• Collaborate with robotics engineers to integrate perception, control, and cognition into a unified system
• Conduct applied research that pushes Physical AI forward, with an emphasis on commercial viability
• Analyze performance, debug complex behaviors, and continuously improve model robustness in the real world

Skills

Machine learning

Robotics/Embodied AI

Python

PyTorch

Publications

Education

MS or PhD in EECS, Robotics, CS, AI, or related field

Tools

PyTorch

TensorFlow

Job description

We're looking for a Physical AI Researcher passionate about redefining the foundation of labor through robotics. You'll help lead the research effort to solve some of the biggest hurdles in robotics: reliability and data efficiency for manipulation. Supported by our humanoid platform that has been shipping for the past year, you will do applied research that brings Physical AI ever closer to mainstream commercial deployment.

About Feather

Feather is a VC-backed startup assembling a world-class team to build general-purpose robots. While competitors pursue end-to-end integration, we are focused on creating the best labor platform on which others will build, monetize, and scale, fostering an ecosystem of solutions akin to Apple’s or Nvidia’s.

We’ve moved through the hardware lifecycle at lightspeed -- building our first prototype in two months, fulfilling our first order in nine months (vs. 31 months for Figure), and working with a $4B revenue manufacturer. Our units are deployed across several fields today, including cooking, retail, manufacturing, and more.

Responsibilities

• Develop, train, and deploy machine learning models for general-purpose robots
• Implement and optimize algorithms in areas such as imitation learning, RL, diffusion policies, VLA, and navigation
• Collaborate with robotics engineers to integrate perception, control, and cognition into a unified system
• Conduct applied research that pushes Physical AI forward, with an emphasis on commercial viability
• Analyze performance, debug complex behaviors, and continuously improve model robustness in the real world

Qualifications

• MS or PhD in EECS, Robotics, CS, AI, or a related field, or equivalent industry experience
• Strong background in machine learning, with applied work in robotics or embodied AI
• Experience with one or more of: Vision-Language Models (VLAs), imitation learning, diffusion policies, reinforcement learning, navigation
• Proficiency in Python and modern ML frameworks (e.g., PyTorch, TensorFlow)
• Track record of impactful projects, whether in academia, open-source, or industry
• Publications in top robotics conferences (ICRA, CoRL, RSS, IROS, etc.)

Why Join Us

• Own the whole research stack and be one of the principal architects of our humanoid platform instead of only working on a small fragment
• Shape the future of robotics and physical AI in a tight-knit, driven team that values ownership, speed, and creativity
• We have the capital, the hardware ownership, and the generational upside to make this your life's work