Embodied-AI Engineer (R&D)

Two tracks: build the robot, or make an existing robot do real work.

Two tracks — pick one first

Build trackApplication track
In one lineBuild the robotMake an existing robot do real work
Start fromOpen-source body: structure, drives, control board, firmwareCommercial body + vendor SDK + demos
Core skillsMechanical, motors and drives, motion control, embeddedSDK/ROS 2, calibration, upper-layer algorithms, integration
Typical coursesMechanics, motor drives, embedded, motion control, simulationROS 2 apps, SLAM and navigation, vision and grasping, VLA
OutputA prototype that stands, walks and teleoperatesA working scenario task (code, parameters, calibration files)
RolesStructure / drive / motion control / embeddedRobot application development, algorithms, integration
EmployersRobot-body companies, research institutes, hardware startupsApplication companies, integrators, end users

What this engineer does

  • Build track: replicate an open-source body → remix structure and drives → reliability, cost and certification → producible product
  • Application track: replicate vendor demos → change scenario and add algorithms → mapping, navigation, recognition, grasping/inspection → delivery and operations

Both tracks converge at productisation.

Where you go

Robot-body companies and hardware startups (build track); application companies, integrators and end users (application track); or turn a product into something sellable at Chuangsheng Works.

How you earn

Salary; licensing/share/equity for signed products (agreed by track).

Who it suits

  • Mechanical/electronics background → build track
  • Software/automation background → application track

Want to try this track? Send us your background for an assessment.