Embodied AI Engineer

About the Role

We are buildingin-house humanoid robotics capabilities (perceive, reason, and act autonomously)for real-world service and operational applications.

Weare looking for a hands-on Embodied AI Engineer with experience in robotlearning to develop, fine-tune, deploy, and validate intelligent behaviors onphysical humanoid robots.

Thisis a startup-style role with broad ownership across AI development, datacollection, robot integration, testing, failure analysis, and continuousperformance improvement. The ideal candidate is comfortable working acrossdisciplines and taking ideas from early experimentation through to real-robotdeployment.

Key Responsibilities

- Develop, integrate, and fine-tune robotfoundation models for perception, mapping, task planning, and embodieddecision-making.

- Own the development cycle from datacollection and model training to physical-robot deployment and validation.

- Test and analyze system failures, thenimprove performance through targeted data collection, model retraining, anditerative validation in both simulation and real-world environments.

- Track key performance metrics, includingtask success rate, completion time, intervention rate, repeatability, andsafety.

Minimum Requirements

- Good knowledge of VLA models, imitationlearning, reinforcement learning, or diffusion policies.
- Master’s degree or above in Robotics, Computer Science, AI, Engineering, or arelated field.

- Willingness to work across AI,robotics, hardware, data collection, testing, and operations.

- Hands-on experience with robotic systems,including integration, calibration, or teleoperation.

Strongly Preferred

Hands-on experience deploying andvalidating AI models on physical robots is a major advantage.

- Proven experience deploying learnedpolicies on real robotic systems, including humanoid, bimanual, dexterous, orcontact-rich manipulation.

- Experience across data collection, modeltraining, deployment, and real-robot validation.

- Experience training andfine-tuning large robot-learning models using cloud-based GPU infrastructure.

- Demonstrable hands-on work throughdeployed systems, projects, publications, or robot demonstration videos.

What We Offer

- The opportunity to workdirectly on embodied AI models and deploy them on physical humanoid robots.

- A hands-on role combining research, modeldevelopment, system integration, and real-world deployment.

- The chance to see your work move beyondsimulation and operate in real environments.

WorkType : Full-time |On-site