The posting
We are seeking a highly motivated Senior Research Assistant to contribute to an interdisciplinary research project investigating physical reservoir computing for advanced robotic manipulation. The role is well suited to a robotics researcher with strong hands-on capability in robot perception, manipulation, controls, simulation, computer vision, CAD, and experimental system integration. The successful candidate will work across mechanical design, additive manufacturing, sensing, physical experimentation, robot control, and AI-enabled autonomy.
Key responsibilities -
- Integrate experimental grippers with robotic arms, mobile manipulators, or other relevant robot platforms.
- Develop and validate robot-manipulation pipelines involving forward and inverse kinematics, motion sequencing, trajectory generation, grasp planning, and closed-loop pick-and-place control.
- Implement calibration procedures, including camera calibration, hand–eye calibration, coordinate-frame alignment, and workspace characterisation.
- Develop feedback-driven manipulation behaviours that use perception and physical-reservoir outputs to adapt grasping, lifting, regrasping, sorting, placement, and recovery actions.
- Support real-world testing of manipulation systems in controlled laboratory and semi-structured environments, with attention to safety, reproducibility, and failure analysis.
- Develop RGB-D perception pipelines for object detection, segmentation, pose estimation, object retrieval, and grasp-relevant geometric reasoning.
- Apply computer-vision and deep-learning methods using PyTorch, OpenCV, and suitable foundation models for perception tasks.
- Build or extend synthetic-data pipelines using CAD assets, headless rendering, Blender, pyrender/EGL, or related tools to generate aligned RGB, depth, segmentation, and pose annotations.
- Develop rigorous evaluation protocols, including held-out test sets, confusion matrices, top-k retrieval measures, grasp-success rates, trajectory metrics, and ablation studies.
- Investigate the use of VLA models and LLMs to translate natural-language instructions and visual scene information into high-level manipulation goals and executable robot actions.
- Integrate VLA/LLM outputs with ROS2-based robot-control frameworks while maintaining deterministic safeguards, state verification, and failure recovery.
Required qualifications -
- Bachelor’s degree in Robotics, Mechanical Engineering, Aerospace Engineering, Electrical Engineering, Computer Science, Mechatronics, or a closely related discipline.
- A postgraduate qualification, or current/recent graduate-level training in Robotics, Automation, AI, Control Systems, or related fields, will be advantageous.
- Demonstrated experience developing robotics, automation, computer-vision, control, or mechatronics projects from concept through testing and validation.
- Strong programming ability in Python; working proficiency in MATLAB is highly desirable. Java or other software-development experience is an advantage.
- Practical experience with ROS2 and robotics software stacks, including robot communication, sensor integration, navigation, motion execution, coordinate frames, and debugging.
- Experience in robot manipulation, inverse kinematics, motion sequencing, grasp planning, trajectory control, or closed-loop control.
- Familiarity with classical and modern control approaches, including PID/PI control, state-space modelling, system identification, LQR, feedback control, trajectory tracking, and optimisation-based control.
- Hands-on experience with CAD and 3D design tools, preferably including SOLIDWORKS, Fusion, Blender, or equivalent platforms.
- Familiarity with simulation environments such as Gazebo, PyBullet, Blender, or similar tools.
- Experience with computer vision and machine learning tools, especially PyTorch and OpenCV.
A strong candidate will also demonstrate persistence in debugging complex integrated systems, sound experimental judgement, care with data quality and reproducibility, and enthusiasm for the emerging intersection of additive manufacturing, embodied intelligence, AI, and next-generation robotic manipulation.



