Founding Robot Learning Engineer

Lucé Robotics; Paris; Hybrid; EUR 95,000 - 130,000; Full time

Job summary

Lucé Robotics is seeking a Robotics ML Engineer to teach robots new skills and improve performance across deployments. You will own the manipulation stack—from teleoperation data collection to training, evaluation, and deployment of policies using imitation learning and RL.

You will drive autonomous capabilities from teleoperation to supervised end-to-end autonomy. The role emphasizes real-time model shipping within edge compute budgets, strong PyTorch proficiency, and building data pipelines

Qualifications

• Experience training and deploying imitation-learning or RL policies on real robots.
• Able to ship models that run in real time within an edge compute budget.
• Strong in PyTorch, data pipelines and robot evaluation.
• Background from a robot learning lab or industry team (manipulation, humanoids, mobile manipulation).

Responsibilities

• Own the manipulation stack: collect teleoperation data, train, evaluate, and deploy policies.
• Develop the path to autonomy task by task: teleoperation, assisted autonomy, then supervised end-to-end autonomy.
• Perform RL post-training in simulation, sim-to-real transfer, and distilling models for onboard deployment.
• Build the data flywheel that improves models across tasks and markets.

Skills

Imitation learning

Reinforcement learning

PyTorch

Data pipelines

Robot evaluation

Tools

Edge computing

Job description

You'll teach robots new skills in the field, and make them more capable with every deployment.

What you'll own

• The manipulation stack: collecting teleoperation data, then training, evaluating and deploying policies (imitation learning, fine-tuned VLA/foundation models)
• The path to autonomy task by task: teleoperation, then assisted autonomy, then supervised end-to-end autonomy
• Reinforcement learning post-training in simulation, sim-to-real transfer, and distilling models to run on the robot's onboard computer
• The data flywheel that turns every field deployment into better models across tasks and markets

Profile

• Has trained and deployed imitation-learning or RL policies on real robots, in the field
• Has shipped models that run in real time within an edge compute budget
• Strong in PyTorch, data pipelines and robot evaluation, reads papers and builds from them
• Background from a robot learning lab or industry team (manipulation, humanoids, mobile manipulation)

What we look for

Real work over credentials. Show us what you have built: projects, a portfolio, open source, what you shipped in past roles. That's what we look at first.