Embodied Intelligence Software Engineer

You will develop and integrate embodied-intelligence algorithms and models into stable robot product features. You will tune algorithms for scenarios such as homes and factories, deploy models on embedded hardware, and optimize software architecture, latency, and system responsiveness. You will build data collection, storage, cleaning, annotation, and automated feedback-loop systems. You will design fault handling and evaluation systems, conduct physical-robot testing, and resolve system issues.

Responsibilities

• Develop and integrate embodied-intelligence algorithms and models into robot products
• Tune algorithms for specific application scenarios
• Deploy and optimize models on embedded hardware
• Optimize robot software architecture, latency, and perception-decision-execution responsiveness
• Build software systems for data collection, storage, cleaning, and annotation
• Build automated data feedback-loop systems for continuous model improvement
• Design fault-handling mechanisms for algorithm and sensor failures
• Establish testing and evaluation systems through physical-robot testing and long-duration operation

Requirements

• Proficiency in C++ and Python
• Software architecture design skills
• Knowledge of computer architecture and operating-system principles
• Knowledge of TCP, UDP, HTTP, WebRTC, and DDS
• Knowledge of TensorRT and ONNX Runtime
• Experience with model quantization and pruning
• Knowledge of deep learning, computer vision, and large language models
• Knowledge of robot sensors, multisensor fusion, kinematics, and dynamics