The aim of the Robotics Lab at DeepMind is to endow robots with the ability to learn how to perform complex manipulation and locomotion tasks. At DeepMind, experts in different scientific fields (e.g. quantum chemistry, neuroscience, game theory) collaborate to solve grand challenges that would enable multiple scientific breakthroughs. The lab is made up of a group of research scientists, research engineers and software engineers specialised in Robotics at DeepMind.
About the role
Our robotics research team focuses on machine learning for robotics, including but not limited to deep reinforcement learning algorithms. The robotics lab collaborates with researchers in machine learning to develop powerful general-purpose learning algorithms for real-world robotics. We’re looking for robotics engineers to work on the problem of learning real robotics tasks with machine learning methods. You should be passionate about the field of robotics and have experience relevant to this field, such as: modelling and planning for manipulation and whole-body motion, real-time motion control with online optimisation methods (e.g. QP, MPC, RL), perception for robotics (e.g. semantic segmentation, scene understanding), and state estimation, etc. The work that we do presents interesting engineering challenges and as a member of the Research Engineering team, you’ll help enhance the performance of research infrastructure and learning algorithms.
- Provide software design and programming support to research projects.
- Report and present software developments including status and results clearly.
- Architect and implement software libraries for robotics applications.
- Collaborate with other researchers to implement and evaluate algorithms.
- Design of robotics experiments and scientific evaluation of results.
- MEng/MSc/PhD degree in engineering, mathematics, physics, or equivalent.
- Good knowledge of either Python or C++.
- Knowledge of machine learning and/or statistics.
- Modelling, planning and control for hybrid dynamical systems (i.e. systems that exhibits both continuous and discrete dynamic behaviours) with application to either manipulation or locomotion.
- Reinforcement learning applied on simulated and real robots (with continuous action space and continuous state space)
- Sensor fusion for state estimation with different sensors: vision, depth sensors, inertial sensors, tactile sensors, proprioception, etc
- Real-time control with on-line optimisation such as MPC (model predictive control) and QP (quadratic programming)
- Robotics perception (object pose estimation and matching, scene segmentation and dense prediction, grasp detection and planning, end-to-end and self-supervised grasping, learning of depth and ego-motion)
- Experience with ROS or comparable middleware
- Working knowledge of Tensorflow or similar frameworks
- Experience with numerical methods and data visualisation
- Experience on deep learning for perception
- Contributions to open source and/or large-scale software projects
- Experience with low-level embedded software and firmware
- Experience in real-time middleware for robotics
- Experience with mechanical and electronic design
DeepMind welcomes applications from all sections of society. We are committed to equal employment opportunity regardless of race, religion or belief, ethnic or national origin, disability, age, citizenship, marital, domestic or civil partnership status, sexual orientation, gender identity or any other basis as protected by applicable law. If you have a disability or additional need that requires an accommodation, please do not hesitate to let us know.
|Job Category||Computing platforms for robotics, engineering, Innovation|