Robot Learning Engineer

Stack

Responsibilities

• ▹Write and enforce the capture spec: cameras, mounts, calibration and what disqualifies an episode
• ▹Own ground truth: measure hand pose and close the gap in warehouse conditions (work gloves, occlusion, cold halls)
• ▹Build the evaluation harness that decides which hours of footage make the corpus
• ▹Post-train open robot-learning models on the company's data and run the first task evaluations
• ▹Follow the literature, reproduce the claims the project depends on and steer capture before money is spent

Requirements

• ▹PyTorch and computer vision with real 3D geometry: camera calibration, pose estimation, SLAM fundamentals
• ▹Ability to read and reproduce papers
• ▹Familiarity with the current robot-learning stack and policy classes
• ▹Experience training on self-collected data and knowledge of training-time failure modes
• ▹Comfortable as the only ML voice: make the call, write it down and revisit it
• ▹Directing agents (Claude Code, Devin) for routine work while keeping the quality bar

Nice to have

• ▹Hand-pose estimation or egocentric video in production
• ▹Multi-camera rigs and time sync
• ▹Experience post-training robot-learning models
• ▹ROS 2

Soft skills

What we offer

• ▹The physical-AI side of the autonomous grocer, with the capture spec and first training runs yours from day one
• ▹A data asset few teams in the world get to build, in fresh groceries
• ▹Off-the-shelf hardware, a managed cloud estate and a mandate that has already started

About the company

Rohlik is the leading Central European e-grocer (Rohlik.cz, Knuspr.de, Kifli.hu, Gurkerl.at, Sezamo.ro) with over a million customers and 17,000 items.