Physical AI
Also written as Embodied AI, Robot Foundation Model, Vision-Language-Action Model, VLA, World Model, World Models
AI that controls machines in the physical world, such as robot arms, humanoid robots, drones and self-driving vehicles, using large models that turn camera input and instructions directly into movement. World models, which simulate how the physical world behaves, are used to train and test these systems.
Think of it like
The difference between a chess engine and someone who can actually pick up the pieces.
Junior or senior?
A small, fast-growing and heavily contested talent pool, mostly from robotics labs, autonomous-vehicle companies and AI research.
Senior sounds like
Has run models on real hardware, not only in simulation, and can talk about the gap between the two.
Ask them
“What worked in simulation but failed on the real robot, and how did you close the gap?”
Sounds like real experience
Names the robot or vehicle and its sensors, and describes a concrete simulation-to-reality failure such as lighting, friction or timing.
Probe further if
Has only worked in simulation or with a hobby kit but describes it as production robotics.