Senior Robotics Research Engineer
Poland Full-time Senior AI Research
At a glance
Senior AI Research role at Botz. Poland · full-time.
Pay not stated Research Engineer salary
RoleAI Research
SenioritySenior
LocationPoland
EmploymentFull-time
PostedOct 16, 2026 · 6d ago
Role brief
aimlroles summary, based on the employer's posting.
What you'll do
- Build simulators for difficult-to-model robotic systems and develop learning and control software used across the robots
- Train learned policies to work on physical robots, bridging simulation and hardware
- Combine learned methods with model-based control, choosing the approach that fits each task
- Move soft robots from initial simulation toward reliable real-world movement and grasping
- Create experiment tools for the wider team to use
What you bring
- At least five years in professional robotics software, or three years after a relevant PhD
- You have trained learned policies and run them on physical robots, not only in simulation
- You have moved a policy from simulation to hardware and can explain where the gap showed up
- You can work fluently in PyTorch or JAX, with GPU-parallel simulation, and write modern C++ and Python
- You know control theory, have used ROS 2 on real projects, and are fluent in English
Who this fits
This senior role suits someone strong in robot learning or classical control who wants to help shape simulation, training and control for soft robots. The lab is near Rondo Rataje in Poznań, and the role includes occasional travel to London and Milan.
From the employer
We’re building something that doesn’t exist yet. Our startup, backed by teams in London and Milan, is working at the frontier of soft robotics. We’re looking for a Senior Robotics Research Engineer who lives at the intersection of machine learning, control theory, and software, and who wants to teach strange, compliant machines to move, and grasp exactly the way we intend.
Our lab is 20 minutes from Rondo Rataje, Poznań.
www.morph.inc
You’ll own the simulation, learning, and control stack that every one of our robots runs on. That means building simulators for systems that are hard to model, training policies that survive contact with the real world, and combining learned and model-based control where each works best. You’ll take soft bodies from first sim to robust real-world behaviour, and build the tools the rest of the team relies on to run experiments. Occasional travel to London and Milan comes with the role.
Potential matters more than a perfect CV.
If you’re strong in robot learning or classical control but still growing in the other, and you want to work on robots that haven’t been built before, apply anyway.
Our lab is 20 minutes from Rondo Rataje, Poznań.
www.morph.inc
You’ll own the simulation, learning, and control stack that every one of our robots runs on. That means building simulators for systems that are hard to model, training policies that survive contact with the real world, and combining learned and model-based control where each works best. You’ll take soft bodies from first sim to robust real-world behaviour, and build the tools the rest of the team relies on to run experiments. Occasional travel to London and Milan comes with the role.
- Have 5+ years of professional robotics software experience, or 3+ years after a PhD in Robotics, Machine Learning for Robotics, Controls, or a related field
- Have trained and deployed learned policies (reinforcement learning or imitation learning) on physical robots, not just in simulation
- Have taken a policy from simulation to hardware, and can tell us exactly where the reality gap bit you
- Are fluent in PyTorch or JAX, and comfortable running large-scale training in GPU-parallel simulators (Isaac Lab, MuJoCo/MJX, Brax, or similar)
- Know your control theory: state-space methods, LQR or MPC, state estimation, and when a classical controller beats a learned one
- Write strong modern C++ and clean Python that other engineers can build on
- Have used ROS 2 on real projects, not just tutorials
- Care about testing, code review, and reproducible experiments, because your code is the foundation others stand on
- Are fluent in English
- Experience with robotic manipulation: dexterous grasping, in-hand manipulation, or contact-rich tasks
- Experience with mobile robotics
- Force, impedance, or tactile-feedback control
- Experience modelling soft, compliant, or continuum systems (Cosserat rods, FEM, SOFA, differentiable simulation)
- System identification, actuator modelling, or domain randomisation for sim-to-real
- Teleoperation and demonstration data pipelines, or experience with vision-language-action and robot foundation models
- Embedded and real-time experience (STM32, RTOS, CAN)
Potential matters more than a perfect CV.
If you’re strong in robot learning or classical control but still growing in the other, and you want to work on robots that haven’t been built before, apply anyway.