Job Description

We are looking for a Research Engineer to drive the “intelligence” of a next-generation construction robot. 

Leading Track 1A of the project, you will develop enhanced environmental perception capabilities, allowing a modular robot to navigate unstructured construction sites, recognize temporary objects, and perform high-precision panel installations. 

There are opportunities to apply state-of-the-art Embodied AI, including Vision-Language-Action (VLA) models and Diffusion Policies, to real-world robotic manipulation.

Key Responsibilities:
•    Develop and implement multi-modal perception pipelines, including BIM-based SLAM, traversability map analysis, and 3D LiDAR/vision sensor fusion.
•    Design few-shot learning algorithms for robust detection of temporary construction objects and obstacles.
•    Develop real-time vision measurement systems and intuitive human-machine interaction protocols (e.g., hand/body pose control) for 6-DoF panel alignment.
•    Train and deploy cutting-edge manipulation policies (e.g., VLA models) utilizing high-performance compute clusters.

Qualifications

•    Min Bachelor’s degree in Computer Science, AI, Robotics, or a related discipline.
•    Expertise in Computer Vision, 3D Point Cloud processing, and Deep Learning frameworks.
•    Solid foundation in 3D geometry, visual SLAM, and generative AI methods applied to robotic manipulation.
•    Open to Fixed Term Contract.