Job Description
The Senior Robotics Software Engineer – AI-Enabled Manipulation & Industrial Automation designs, develops, integrates, validates, and deploys AI-enabled robotic applications for industrial and factory automation. The role spans robotic manipulation, perception, motion planning, computer vision, simulation, controls, and AI—including vision-language-action (VLA) models and learned manipulation policies.
The engineer develops end-to-end, closed-loop solutions that combine classical robotics and modern AI for tasks such as perception-guided manipulation, multimodal task understanding, sensor-driven feedback, and autonomous execution. This role bridges robotics software, artificial intelligence, and manufacturing automation, with a focus on turning advanced prototypes into robust, scalable, production-relevant systems.
Key responsibilities
Design and deploy robotic applications for industrial automation, including pick-and-place, material handling, part transfer, machine tending, inspection support, and vision-guided manipulation.
Build collision-aware, dynamically feasible manipulation pipelines integrating object detection and pose estimation, grasp generation, inverse kinematics, motion and trajectory planning, collision avoidance, reliable execution, and gripper control using frameworks such as MoveIt, Lula, RMPflow, OMPL, or related technologies.
Develop computer-vision and AI perception systems for detection, semantic and instance segmentation, 6D pose estimation, RGB-D and point-cloud processing, scene understanding, and grasp-point generation.
Integrate AI/ML models into real-time or near-real-time robotic workflows to improve perception, generalization, manipulation, task-level decision-making, and runtime reliability.
Develop closed-loop robotic systems that use sensor feedback to monitor task execution, detect failures, update robot actions, and improve runtime reliability.
Develop and evaluate VLA models and learned policies for multimodal task understanding, natural-language instruction following, action generation, imitation learning, and autonomous manipulation.
Use NVIDIA Isaac Sim, Isaac Lab, Omniverse, Replicator, and related tools to build digital environments, generate synthetic data, validate behavior, and develop simulation-to-real workflows before physical deployment.
Own robotic software architecture across perception, planning, control, simulation, AI inference, data flow, system interfaces, hardware integration, and deployment; develop modular, maintainable software using Python, C++, ROS 2, Linux, Git, Docker, and reproducible development environments.
Integrate and calibrate manipulators, grippers, RGB/RGB-D and depth cameras, force/torque sensors, and other automation hardware, including camera intrinsics/extrinsics, hand-eye calibration, and robot-to-cell registration; make calibration robust and repeatable in the field.
Implement force and compliant control—such as impedance, admittance, and guarded moves—for contact-rich tasks including insertion and assembly.
Build data-collection workflows for learned policies, including teleoperation, demonstration capture, dataset curation, and experiment tracking.
Debug and validate systems across software, AI models, perception, simulation, motion planning, controls, networking, and hardware. Define measures such as task success, perception accuracy, cycle time, trajectory quality, repeatability, and reliability.
Collaborate with robotics, perception, ML, controls, manufacturing, and plant-facing teams to troubleshoot issues, document solutions, and transition prototypes from simulation through hardware integration, validation, and deployment.
Provide technical ownership of robotic applications from early prototyping and simulation through hardware integration, validation, and deployment.
Optional: Integrate robotic cells with industrial systems and safety infrastructure, including PLCs (OPC UA, Ethernet/IP, Modbus), safety I/O, and robot…