Join Muguang

Own the mechanical design and engineering of the mobile base, telescopic structure, and manipulator body.

1. Design the mobile base, telescopic mechanisms, manipulator links, and joints.
2. Validate interference, range of motion, payload, strength, stiffness, and assembly space.
3. Maintain the BOM, coordinate with suppliers, and resolve manufacturing and integration issues.

Develop dexterous hands and end effectors for household grasping and delivery tasks.

1. Select and design underactuated multi-finger hands, parallel grippers, or compliant end effectors.
2. Integrate actuation, force, tactile, or pressure sensing and develop ROS2 drivers or SDKs.
3. Design grasping experiments and document success rates and failure causes.

Own the robot's electrical architecture, power management, motor-control hardware, sensor interfaces, and wiring system.

1. Design 24V/48V power distribution, protection, grounding, and electrical architecture.
2. Design or select motor-control boards and complete schematics, PCB layouts, and sensor-interface integration.
3. Handle wiring, EMI/EMC, thermal, and hardware debugging tasks while maintaining the hardware BOM.

Develop MCU/RTOS firmware, communication protocols, ROS2 hardware drivers, and low-level safety mechanisms.

1. Develop controller and motor-driver firmware, including scheduling, communications, and fault recovery.
2. Implement motor and joint-torque control, safety protection, and CAN, EtherCAT, and RS485 communications.
3. Build ROS2 drivers for the base, odometry, IMU, and manipulator joints.

Develop autonomous navigation, obstacle avoidance, localization, mapping, and real-robot deployment for mobile bases.

1. Develop global planning, local planning, and dynamic-obstacle handling with ROS2 and Nav2.
2. Model the mobile base and tune velocity constraints, controllers, and navigation parameters.
3. Integrate localization, mapping, semantic maps, and sensor data, then validate on the real robot.

Develop manipulator kinematics, trajectory planning, collision checking, force control, and contact-rich manipulation.

1. Implement forward/inverse kinematics, joint and Cartesian trajectory planning, and collision checking with MoveIt2.
2. Develop compliant and force-control capabilities, including impedance and admittance control.
3. Integrate end effectors, perception, and task-planning modules and improve real-robot trajectory stability.

Develop 3D scene understanding, object detection, pose estimation, localization, mapping, and production-ready perception deployment.

1. Develop RGB-D point-cloud processing, object detection, and 6DoF pose estimation.
2. Integrate visual or LiDAR SLAM and build semantic maps for object-search tasks.
3. Optimize edge inference and establish benchmarks for accuracy, recall, and real-time performance.

Build natural-language task understanding, high-level planning, and interfaces across perception, planning, and control.

1. Select and deploy VLM/LLM solutions for home scenarios, task understanding, and task decomposition.
2. Design behavior trees, task graphs, or state machines with exception handling and failure recovery.
3. Build evaluations for grasping, delivery, switching, and tidying tasks and close the execution loop.

Develop robot manipulation policies, demonstration-data pipelines, model training, and real-robot evaluation.

1. Build data, training, evaluation, and real-robot validation pipelines for imitation or reinforcement learning.
2. Establish teleoperation data-collection workflows for household manipulation demonstrations.
3. Train grasping, delivery, switching, and tidying policies and use failure cases to drive the data flywheel.

Own system-level functional, stability, safety, and care-scenario testing for complete robot prototypes.

1. Define system test plans covering functionality, endurance, safety, and critical household tasks.
2. Build automated regression tests using ROS2 bag playback, simulation, and test scripts.
3. Establish task success, collision, and recovery metrics and drive defects to closure.

Develop ROS2 application nodes, robot-data pipelines, simulation utilities, and internal engineering tools.

1. Develop ROS2 application nodes for task scheduling, state machines, and human-robot interaction.
2. Build rosbag2 recording, cleaning, conversion, and training-platform upload pipelines.
3. Create calibration, log-analysis, and performance-evaluation tools and maintain internal Python SDKs.

Develop cloud services for robot device management, remote monitoring, OTA updates, logs, and business APIs.

1. Develop services for device registration, status reporting, remote commands, and alert delivery.
2. Implement firmware/model OTA, version management, staged rollout, and rollback capabilities.
3. Build log and operational-data services and RESTful APIs while owning deployment, security, and reliability.

Own model-training infrastructure, distributed training, experiment management, evaluation, and edge deployment.

1. Build data-loading, preprocessing, training, validation, and checkpoint-management pipelines.
2. Optimize distributed training and GPU utilization and establish reproducible experiment and evaluation workflows.
3. Deliver model quantization, inference optimization, and deployment to edge devices such as Jetson.

Own complete-robot bring-up, cross-module interfaces, the system software framework, and system-level integration.

1. Integrate mechanical, electrical, embedded, perception, motion-control, and intelligent-model modules.
2. Define ROS2 messages, coordinate frames, timing, synchronization, and fault-propagation boundaries.
3. Establish repeatable build, release, deployment, and regression environments and diagnose system defects.

Build frontend experiences for robot administration, family and care-organization users, and operations teams.

1. Build administration interfaces for device status, task history, alerts, logs, and OTA management.
2. Develop family and care-organization interfaces for status viewing, service records, notifications, and settings.
3. Integrate RESTful APIs and WebSocket/SSE updates across desktop, tablet, and mobile devices.

Own robot-operation data collection, cleaning, annotation, quality control, and dataset version management.

1. Collect teleoperation, real-robot, and simulation data for priority home-care tasks.
2. Define annotation guidelines and taxonomies and coordinate annotation, review, and consistency checks.
3. Clean, align, quality-check, and version datasets to support training and failure-case feedback loops.