Senior AI Engineer, Dexterous Manipulation & RL

Amazon; Zürich; On-site; CHF 150,000 - 230,000; Full time

Job summary

RIVR, an Amazon company, is building Physical AI with autonomous robots for real-world doorstep delivery. This Senior AI Engineer role focuses on reinforcement learning, deep learning, and dexterous manipulation, using both simulated and real-world data to address dynamic grasping and object interaction.

You will develop cutting-edge RL algorithms, work with the foundation model team, and translate sensor input into precise motor commands, enabling reliable manipulation across diverse packages

Qualifications

• Master's degree or higher in Engineering, Robotics, or Machine Learning.
• At least five years of industry or research experience; PhD experience applicable.
• Strong deep learning and reinforcement learning fundamentals, including MDPs and related topics.
• Background in robotics, autonomy and/or manipulation.
• Experience deploying neural networks on hardware platforms and writing production-level code in C++.
• Prototyping algorithms and training deep neural networks in Python.

Responsibilities

• Develop cutting-edge reinforcement learning algorithms for robust, contact-rich dexterous manipulation using multimodal sensor input.
• Design, test, and refine algorithms to solve real-world manipulation challenges such as diverse package forms and door-hand-offs.
• Collaborate with the foundation model team to leverage both simulated and real-world data.

Skills

Reinforcement learning

Deep learning

Python

C++

Robotics

MDP

Sim-to-real

Transfer learning

Dexterous manipulation

Education

Master's degree or higher in Engineering, Robotics, or ML

PhD preferred in Robotics/CS/ML

Tools

Python

C++

Job description

RIVR, an Amazon company, is building Physical AI with autonomous robots for real-world doorstep delivery. This Senior AI Engineer role focuses on reinforcement learning, deep learning, and dexterous manipulation, using both simulated and real-world data to address dynamic grasping and object interaction.

You will develop cutting-edge RL algorithms, work with the foundation model team, and translate sensor input into precise motor commands, enabling reliable manipulation across diverse packages