Vision Engineer, Manipulation Policy Learning

Vedi l’offerta su Go Lavoro: golavoro.it/annuncio-lavoro/8o5mtk

About the Company
https://www.aicamrobotics.ai/ We are a forward-thinking company that is developing autonomous robotic solutions for last mile delivery and intelligent devices for smart mobility.

About the Role
We are looking for a Vision Engineer (f-m-d) to join the team teaching our robots to solve manipulation tasks in real-world hospitality environments. Your focus is policy learning and the data behind it: generating and curating demonstrations, training visuomotor policies, and getting them to run reliably on the real robots. You will work closely with the robotics engineers who own the teleoperation, motion planning and control pipelines. The goal is policies that generalize across venues and robot embodiments, and scale from short skills to long-horizon activities such as stripping a bed or tidying a room. You will work at the frontier of imitation learning, vision-language-action models and large-scale data generation, on real robots. Responsibilities
• Train, benchmark and deploy visuomotor policies (e.g. ACT, Diffusion Policy, VLA-class models) on real hospitality tasks, with clear success metrics and evaluation protocols; iterate on failures through data, retraining and redeployment.
• Design observation and action spaces that transfer across different robot embodiments.
• Scale and improve data collection: shape the teleoperation setup run by the robotics engineers (task design, camera placement, synchronization, data format), curate the resulting datasets, and go beyond teleoperation with demonstration augmentation and generation (MimicGen, SkillMimicGen, Real2Render2Real), simulation, and real-to-sim pipelines.
• Develop the perception that supports manipulation: object detection and segmentation,



pose estimation, and foundation vision models used to condition policies.
• Work with the motion planning and control engineers to run policies on the arm control stack (end-effector-space control, safety limits).
• Make manipulation safe around people and fragile objects. Required Skills
• Strong background in deep learning for computer vision or robot learning.
• Experience with training and deploying imitation learning policies (ACT, Diffusion Policy, or VLA models).
• Proficiency in Python and PyTorch, including multi-GPU training.
• Solid understanding of the data side of robot learning: camera calibration, time synchronization, latency, and dataset formats.
• Proven ability to diagnose why a learned policy fails on the robots (distribution shift, camera changes, timing) and fix it. Nice to have Experience in any of the following is a plus. We do not expect one person to cover all of them:
• Hands-on experience on real robots.
• Fine-tuning vision-language-action models (OpenVLA, π0, GR00T) to new embodiments and tasks.
• GPU-accelerated simulation (Isaac Lab, ManiSkill, MuJoCo).
• Foundation vision models (DINOv2, SAM, CLIP) and vision-language models for perception or task conditioning.




• World models and video generation for planning or policy evaluation.
• Reinforcement learning fine-tuning of imitation policies and sim-to-real transfer.
• Model optimization for edge inference (TensorRT, NVIDIA Jetson).
• Publications at top-tier vision or robotics venues (CVPR, ICCV, CoRL, RSS, ICRA). What we are looking for
• Engineers who bridge theory and real-world execution and are comfortable both reading research papers and running experiments on real robots.
• Strong debugging

skills:
able to work out why a policy fails, whether the cause is the data, the model, the cameras or the control loop.
• Comfort with the full learning pipeline, from data collection to training and deploying policies on the robot.
• Genuine interest in solving open robot learning problems rather than applying standard pipelines. Pay range and compensation package
• Bargaining Agreement: Italian National Collective Bargaining Agreement (CCNL Metalmeccanico Industria).
• Classification Target: Expected job level B2, corresponding to a statutory base minimum of € 33058,74 gross per year.
• Total Rewards Package: Final financial offers will exceed collective bargaining minimums to reflect individual expertise and prior achievement.
• The complete rewards proposal will include an individual merit allowance on top of the base minimum, a fixed production allowance, welfare benefits, performance bonuses, healthcare coverage (Metasalute), and pension options (Cometa). Equal Opportunity Statement We are an Equal Opportunity Employer. Applications are welcome from all individuals regardless of gender, age, sexual orientation, ethnicity, background, or disability status.

Vedi l’offerta su Go Lavoro: golavoro.it/annuncio-lavoro/8o5mtk