Dexterous Hand and End-Effector Control Engineer

Job Responsibilities

• Responsible for the design, development, debugging, and optimization of control modules related to robot dexterous hands, grippers, and end effectors.
• Responsible for dexterous hand multi-degree-of-freedom motion control, finger coordination control, grasping control, fingertip contact control, and end-effector action strategy development.
• Responsible for the interface adaptation, coordinate system calibration, kinematic modeling, and control integration of various end effectors such as dexterous hands, grippers, long-handled tools, and soft-padded ends.
• Responsible for the design of end-effector control strategies for robots in tasks such as grasping, clamping, pressing, flipping, rotating, pushing and pulling, lifting, and correcting posture.
• Responsible for closed-loop control based on visual, force, tactile, current, or joint state feedback, improving the stability, compliance, and success rate of end-effector operations.
• Responsible for the coordinated control between the dexterous hand and the robotic arm, connecting the complete operation chain from the robotic arm's pose control to finger movement control.
• Responsible for building dual-hand coordination capabilities, supporting tasks such as dual-hand grasping, dual-hand carrying, stabilizing with one hand while operating with the other, dual-hand lifting, and dual-hand alignment.
• Responsible for the development of force limits, speed limits, collision protection, slip detection, clamping stability assessment, and abnormal recovery strategies during the end-contact process.
• Responsible for end-effector operation debugging, parameter tuning, failure case analysis, and operation strategy iteration in real robot environments.
• Validate end-effector manipulation in simulation and on physical robots, taking tasks such as grasping, opening doors, pressing buttons and switches, and using tools from prototype validation to real-world deployment.

1. Basic General Ability (Base)

• Proficient in Linux/Ubuntu development environments, with experience in robot system development and on-robot debugging.
• Proficient in C and Python, with good skills in code structure design, debugging, and engineering implementation.
• Familiar with the ROS / ROS2 system, understanding common mechanisms such as Topic, Service, Action, TF, parameter management, and rosbag.
• Familiar with the basics of robot kinematics, including forward and inverse kinematics, Jacobian matrix, coordinate transformation, SE(3), quaternions, and pose calculation.
• Understand the coordinate relationship between the robotic arm and the end effector, capable of performing end tool calibration, TCP calibration, and hand-eye relationship handling.
• Familiar with common control methods, including position control, velocity control, force control, impedance control, admittance control, and force-position hybrid control.
• Possess debugging experience with real robots, robotic arms, dexterous hands, grippers, or end tools.
• Able to handle engineering issues during real operations such as grasp failures, target slipping, unstable contact, force control anomalies, and motion limit exceedances.
• Proficient in using Git, with good coding standards, documentation habits, testing awareness, and the ability to collaborate across modules.

II. Specialized Technical Directions

Track A: Dexterous-Hand Control and End-Effector Development

• Familiar with the control architecture of dexterous hands, grippers, multi-fingered hands, or custom end effectors.
• Understand issues related to finger kinematics, finger coordination, joint limits, contact constraints, and grasp stability.
• Capable of completing dexterous hand drive interfaces, controller packaging, motion interface design, and state feedback processing.
• Able to design finger movement strategies according to different tasks, such as enveloping grasp, two-finger pinch, three-finger hold, palm pushing, fingertip pressing, etc.
• Experience with dexterous hands, grippers, haptic gloves, end tools, or self-developed actuator debugging is preferred.

Track B: Grasping, Contact, and Fine-Motion Control

• Familiar with robot grasp planning, grip force control, contact detection, slip detection, and grasp stability assessment.
• Capable of combining visual, depth, point cloud, or pose estimation results to generate end-effector grasp points, approach directions, and finger closure strategies.
• Able to handle the differences in grasping irregular objects, small objects, lightweight objects, rigid objects, and flexible objects.
• Familiar with fine motor tasks such as pressing buttons, toggling switches, turning knobs, grasping door handles, and lifting lids.
• Capable of determining contact status and performing online correction based on force sensation, tactile feedback, current feedback, or joint torque estimation.
• Experience with tactile sensors, six-axis force sensors, flexible fingers, soft grippers, or force-controlled grasping is preferred.

Track C: Bimanual Coordination and Complex Manipulation

• Familiar with the coordinate relationships, task decomposition, synchronous control, and collision constraints in dual-arm or dual-hand collaborative operations.
• Able to design coordinated actions such as one hand fixing while the other operates, one hand lifting while the other adjusts, both hands grasping, both hands carrying, and both hands correcting.
• Understand the issues of maintaining relative posture, force distribution, synchronized trajectory, target stability, and self-collision avoidance in dual-hand operations.
• Able to cooperate with higher-level task planning, encapsulating bimanual coordinated actions into reusable operational skills.
• Experience with dual-arm robots, humanoid robots, mobile manipulation robots, or dual-hand dexterous manipulation is preferred.

Track D: End-Effector and Task-Skill Library

• Able to establish kinematic models and control interfaces for different end-effectors, such as long-handled grippers, soft-pad door-pushing tools, suction cups, claws, and dexterous hands.
• Able to encapsulate end-effector actions into reusable skills, including grasping, placing, pushing together, pulling, lifting, upright positioning, pressing, toggling, and rotating.
• Able to iterate the end control strategy based on task execution results and failure samples.
• Able to collaborate with perception, SLAM, motion control, and task planning modules to complete closed-loop operations from target recognition to end execution.
• Experience with action primitive libraries, skill libraries, teleoperation teaching, or imitation learning data collection is preferred.

Preferred Qualifications

• Practical project experience with dexterous hands, multi-fingered hands, grippers, robotic arm end-effectors, or tactile sensors.
• Experience with humanoid robots, dual-arm robots, mobile robotic arms, service robots, or industrial robotic arm operation projects.
• Familiar with common dexterous hands or gripper hardware, such as Allegro Hand, Shadow Hand, Inspire Hand, DexHand, Robotiq, Franka Hand, etc.
• Familiar with tools or frameworks such as MoveIt, Pinocchio, Drake, MuJoCo, Isaac Sim, Gazebo, KDL, RBDL.
• Has experience in visual servoing, haptic feedback, force-controlled grasping, slip detection, and contact state estimation.
• Experience with real operational tasks such as pressing buttons, turning knobs, toggling switches, operating doors, handling tools, righting objects, and carrying items with both hands.
• Has experience in teleoperation, teaching learning, imitation learning, reinforcement learning, or operational data collection.
• Experience in high-frequency control, real-time systems, CAN/CAN-FD, EtherCAT, serial ports, motor drives, and actuator debugging.
• Have experience in robot competitions, RoboMaster, automation competitions, or as a core member of complex software and hardware system projects.
• There are high-quality open-source projects on GitHub with a good sense of engineering design aesthetics.
• Experience in AI-First development, familiar with tools like Claude Code / Codex to improve R&D efficiency.