Senior AI Engineer - Dexterous Robotic Manipulation & RL

Amazon Science; Zürich; On-site; CHF 140,000 - 190,000; Full time

Job summary

RIVR, an Amazon company, is advancing Physical AI by deploying autonomous robots for doorstep delivery. We seek a Senior AI Engineer focused on reinforcement learning and deep learning to enable dexterous manipulation using vision, depth, tactile, and proprioceptive inputs.

You will work on both simulated and real-world data to solve dynamic grasping and manipulation tasks. You will collaborate across teams to translate research into production-ready algorithms and contribute to the safety and

Qualifications

• Master’s or higher in a relevant field such as Engineering, Robotics, or ML.
• 5+ years industry/research experience; PhD experience applicable.
• Strong DL fundamentals and RL including MDPs, network architectures, policy optimization, sim-to-real transfer.
• Robotics background with autonomy/manipulation.
• Experience deploying neural networks on hardware.
• Proficient production C++ and prototyping in Python.
• Expertise in dexterous manipulation, grasp synthesis, and trajectory optimization.

Responsibilities

• Develop cutting-edge RL algorithms for robust dexterous manipulation using multi-modal sensors.
• Address real-world manipulation challenges like dynamic hand-offs and door handles.
• Collaborate with foundation model team to leverage simulated and real-world data.

Skills

Deep learning

Reinforcement learning

MDPs

C++

Python

Robotics manipulation

Education

Master's degree or higher

PhD advantageous

Job description

RIVR, an Amazon company, is advancing Physical AI by deploying autonomous robots for doorstep delivery. We seek a Senior AI Engineer focused on reinforcement learning and deep learning to enable dexterous manipulation using vision, depth, tactile, and proprioceptive inputs.

You will work on both simulated and real-world data to solve dynamic grasping and manipulation tasks. You will collaborate across teams to translate research into production-ready algorithms and contribute to the safety and