Job Description
We are looking for a Robotics Engineer specialized in Reinforcement Learning (RL) to develop, deploy, and validate control policies on real robotic systems. You will work at the intersection of simulation and hardware, building robust policies in simulation (MuJoCo / MJLab) and transferring them to real robots with high reliability. This role requires strong expertise in policy learning, sim-to-real transfer, and system identification, as well as hands-on experience validating behaviors on physical platforms.
Primary duties:
* Develop RL-based control policies for robotic systems using MuJoCo and MJLab.
* Design and train locomotion and manipulation policies, including walking, balancing, grasping, and object interaction.
* Implement and compare pure reinforcement learning approaches with imitation learning methods like behavior cloning and offline RL.
* Deploy and validate trained policies on real robot hardware to ensure effective performance.
* Bridge the sim-to-real gap using system identification, domain randomization, and actuator/dynamics modeling.
* Analyze failures and iterate on models, reward functions, and training setups to improve performance.
* Integrate policies into existing robotics stacks, such as ROS 2-based systems.
* Collaborate across teams, interfacing with control, perception, and hardware specialists.
Qualifications
Mandatory Requirements:
* MSc or PhD in Robotics, Machine Learning, Control, or a related field.
* Strong experience in Reinforcement Learning specifically applied to the field of robotics.
* Hands-on experience with MuJoCo, with MJLab proficiency considered a strong plus.
* Proven experience deploying learned policies on real robotic systems and hardware.
* Solid understanding of robot dynamics, control, policy optimization (PPO, SAC, etc.), and imitation learning.
* Experience with sim-to-real transfer techniques, such as domain randomization and system identification.
* Proficient programming skills in both Python and C++.
* Familiarity with robotics frameworks, particularly ROS 2.
Other valued skills (not mandatory):
* Experience with humanoid or legged robots
* Knowledge of actuator modeling and transmission dynamics
* Experience designing robust policies under uncertainty
* Familiarity with GPU-based training pipelines
* Background in optimization or model-based control (MPC is a plus)
Additional Information
Job conditions
We offer a competitive compensation package, including salary, benefits, and opportunities for professional development. You will be part of a dynamic and international team in a constantly growing and developing environment in a fulfilling and inclusive equal opportunity workplace.
* Permanent full-time contract. 37.5 hours/week - At PAL, we believe in work-life balance and in order for every one of our employees to benefit, from 2024 we have pledged to work 2.5 hours less every week.
* Flexible working hours and early finish at 14:00 on Fridays.
* 26 working days of annual leave.
* Opportunities for payment in kind and ongoing training initiatives.
* Free coffee & tea provided.
* Immediate start date.
Contact details
If you are interested, please apply via the link on this site or send your CV to recruit@ pal-robotics.com. Please specify the job code JOB-2026-05 and the name of the job in the subject field of the email.
All information will be treated in accordance with the GDPR.