As a Machine Learning Engineer – Simulation & Robotics, you will design and build simulation infrastructure to
accelerate robot learning. Your work will enable rapid iteration on learning-based approaches by reducing
dependence on physical robot evaluations, constructing realistic and scalable simulations, and generating large-scale
training data. You will help bridge the gap between simulated and real robot behaviour, supporting training and
evaluation efforts and pushing the boundaries of embodied AI.
RESPONSIBILITIES
Own key parts of the robot learning pipeline, with responsibility for the performance and reliability of learned behaviours on physical robots.
Develop and train learning-based robotic policies using reinforcement learning, imitation learning, and other relevant approaches.
Build and improve simulation and sim-to-real workflows, including domain randomisation, system identification, and robustness-driven training.
Deploy and iterate on learned policies on physical robots, diagnosing differences between simulated and real-world behaviour.
Work closely with the wider engineering team to integrate learned behaviours into physical robotic systems.
Develop evaluation methods and tooling to measure policy performance, robustness, and real-world behaviour.
QUALIFICATIONS
The areas below reflect the broader scope of the role. We don't expect candidates to have deep experience in all of them, but we look for strong foundations, relevant strengths, and the ability to develop quickly in less familiar areas.
Demonstrated experience building and evaluating learning-based robotics systems. Candidates with 3+ years at this level may be considered for a senior role.
MS or PhD in Computer Science, Robotics, or a closely related field is preferred.
Strong Python and software engineering skills are essential, with experience developing, debugging, and extending machine learning or robotics systems.
Strong foundations and practical experience in machine learning, particularly in areas relevant to robot learning such as reinforcement learning, imitation learning, or policy learning.
Experience developing and training learned policies, ideally for robotic or embodied systems.
Experience in one or more relevant areas such as robotics simulation, sim-to-real, domain randomisation, system identification, or robustness-driven training.
Familiarity with PyTorch or an equivalent machine learning framework.
Strong systems-level debugging and problem-solving ability, with the ability to investigate issues across learning, simulation, software, and physical behaviour.
Demonstrated ability to quickly develop proficiency in unfamiliar technical areas and apply new knowledge to engineering problems.
Experience deploying and iterating learned policies on physical robots would be particularly valuable.
BENEFITS
Work at the frontier of physical AI, with direct access to advanced robot hardware and edge compute.
Significant technical ownership in a small, highly experienced team.
Competitive salary commensurate with experience.
25 working days of holiday per year, in addition to public holidays.
Visa sponsorship.
As a Junior ML Engineer, you'll work on the software and machine learning systems behind robot learning,spanning simulation, training workflows, evaluation and testing on physical robots
RESPONSIBILITIES
Take ownership of well-defined engineering problems, with increasing responsibility as you develop.
Build and iterate on simulation environments for robot learning.
Implement and test reinforcement learning and imitation learning workflows.
Contribute to training pipelines, data generation and evaluation tooling.
Debug and improve simulation and training workflows through testing, profiling and benchmarking.
Work closely with the wider engineering team to test and validate learning-based systems in simulation and on physical robots.Write clean, well-tested Python code as part of the wider robotics software stack.
QUALIFICATIONS
Strong foundational programming skills are essential.
You should be able to design, write, understand and debug code indewpendently, with a strong grasp of core programming concepts.
A strong academic background in Computer Science, Robotics, Machine Learning, Engineering or a closely related technical field.
Strong academic, research or project experience in relevant areas such as machine learning, robotics, reinforcement learning, control or simulation.
Familiarity with machine learning frameworks such as PyTorch, gained through research, coursework, projects or internships.
Strong problem-solving skills, with the ability to learn quickly and apply new knowledge to technical problems.
Hands-on exposure to physical robotics or learning-based robotic systems would be valuable but is not required.
This role is full-time and on site in Cambridge. You must be willing to relocate to Cambridge if required.
BENEFITS
Hands-on experience at the frontier of physical AI, with access to advanced robot hardware and compute.
Mentorship and increasing technical ownership as you develop.
Competitive salary based on skills and experience.
25 working days of holiday per year, in addition to public holidays.
Visa sponsorship
As a Robotics Software Engineer, you’ll develop the software that connects high-level robot behaviours with reliable execution on physical robotic systems.This is a hands-on robotics software engineering role spanning control, planning and real-world robot operation, working across software and physical hardware.
RESPONSIBILITIES
Own key components of the robot control and task execution stack, with responsibility for their performance and reliability on physical robots.
Develop real-time robotics software in C++, including components for control, motion execution and task-level behaviour.
Design and improve control and planning systems that enable robots to execute behaviours reliably and respond to changing conditions.
Diagnose and resolve system-level issues at the intersection of control, planning, software and robot hardware.
Work closely with the wider engineering team to integrate and validate control and task behaviours on physical robots.
Build testing, instrumentation and health-monitoring tools to evaluate behaviour, diagnose failures and improve system reliability.
QUALIFICATIONS
The areas below reflect the broader scope of the role. We don't expect candidates to have deep experience in all of them, but we look for strong foundations, relevant strengths, and the ability to develop quickly in less familiar areas.
Strong C++ programming skills are essential, with experience developing and debugging robotics, real-time, or systems software.
Good foundations in robotics, with experience in robot control and/or high-level task planning and depth in at least one of these areas.
Hands-on experience developing software for physical robotic systems.
Familiarity with robotics software environments and tools such as ROS/ROS2, robot SDKs, real- time Linux, or simulation platforms.
Strong systems-level debugging and problem-solving ability, with demonstrated ability to quickly develop proficiency in unfamiliar technical areas and apply new knowledge to engineering problems.
Familiarity with learning-based control, reinforcement learning or learned robotic policies would be valuable but is not required.
This role is full-time and on site in Cambridge. You must be willing to relocate to Cambridge if required.
BENEFITS
Work at the frontier of physical AI, with direct access to advanced robot hardware and compute.
Significant technical ownership in a small team, with the opportunity to make a direct impact.
Competitive salary commensurate with experience.
25 working days of holiday per year, in addition to public holidays.
Visa sponsorship
As a Software Engineer – GPU & Edge Computing, you’ll develop and optimise the software that enables demanding robotics workloads to run efficiently across GPU and edge compute systems.This is a hands-on role focused on real-time performance, efficiency and reliability, working across software, compute and physical robotic systems.
RESPONSIBILITIES
Own the development and performance of key components across the GPU and edge compute stack.
Develop and implement high-performance software for GPU and edge compute environments, including C++ and CUDA components for compute-intensive robotics workloads.
Optimise robotics workloads, including training, inference and runtime execution, for latency, throughput, memory usage and compute efficiency.
Identify and resolve performance limitations across software, compute, memory, data movement, and hardware.
Integrate GPU and edge compute platforms within end-to-end robotic systems, ensuring reliable performance on physical hardware.
Build and maintain benchmarking, profiling, and performance monitoring tools to measure system performance, track regressions, and validate improvements.
QUALIFICATIONS
The areas below reflect the broader scope of the role. We don't expect candidates to have deep experience in all of them, but we look for strong foundations, relevant strengths, and the ability to develop quickly in less familiar areas.
Strong C++ programming skills are essential, with the ability to build, debug and optimise performance-critical software.
A good foundation in systems and performance engineering, including experience with profiling, optimisation and investigating performance limitations across software, compute, or hardware.
Some experience with GPU programming or GPU-accelerated computing, with an understanding of GPU architecture and an interest in developing further in areas such as CUDA.
Understanding of relevant areas such as parallel computing, memory management or hardware- aware software optimisation.Exposure to Linux-based edge, embedded or resource-constrained compute environments.
Exposure to robotics systems, physical robot hardware or machine learning workloads would be valuable but is not required.
Demonstrated ability to quickly develop proficiency in unfamiliar technical areas and apply new knowledge to engineering problems.
This role is full-time and on site in Cambridge. You must be willing to relocate to Cambridge if required.
BENEFITS
Work at the frontier of physical AI, with direct access to advanced robot hardware and edge compute.
Significant technical ownership in a small team, with the opportunity to make a direct impact.
Competitive salary commensurate with experience.
25 working days of holiday per year, in addition to public holidays.
Visa sponsorship.