Physical Intelligence
AI Frontier · Robotics foundation models (PI)
Physical Intelligence currently has 7 active AI-related job listings. The majority of these roles, 71%, are focused on data. Engineering is the dominant function, with 5 positions. The company is hiring for roles that frequently involve embodied_ai, multimodal, and vision technologies.
AI Frontier · Robotics foundation models (PI)
Currently tracking 7 active AI roles, up 17% versus the prior 4 weeks. Primary focus: Data · Engineering. Salary range $52k.
Physical Intelligence currently has 8 active AI-related roles in our index. The most common open titles are: Applied Researcher, Controls Engineer, Deployments Software Engineer, ML Infra Engineer, ML Infra Engineer (TPU/Jax/Optimization). Most positions are in Engineering and Research.
Physical Intelligence's active AI hiring is concentrated in: data (63%), agents (25%), application (13%). These categories follow a seven-stage AI lifecycle: data, pre-training, post-training, serving infrastructure, agents, evaluation, and application.
Physical Intelligence is hiring AI talent in: United States (8 roles).
Job postings at Physical Intelligence most frequently reference: embodied ai, model serving, multimodal, rl robotics, inference infra.
In the past 30 days, Physical Intelligence has posted 2 new AI-related roles.
| Title | Stage | AI score |
|---|---|---|
| Applied Researcher This role focuses on deploying and debugging learned policies on physical robots, involving training and tuning policies, writing production-quality code, and debugging the full stack. It bridges research and operations by translating research advances into deployable systems and gathering feedback from real-world failures. The role requires hands-on experience with robot deployment, strong engineering skills, and a practical, debugging-oriented mindset. | AgentPost-train | 9 |
| Deployments Software Engineer This role focuses on deploying AI-powered robots into the physical world by building and optimizing the software systems that enable teleoperation, real-time performance, and reliability. It involves integrating foundation models with robot hardware, designing user interfaces for control, and ensuring the system is robust enough for customer operations. | AgentServe | 7 |