Tactile Robotics Scientist — Dexterous Grasp Policy

Amazon Science; Sunnyvale (CA); Sur place; USD 172 000 - 222 000; Plein temps

Résumé du poste

Amazon is seeking talent to develop advanced robotics systems that combine AI, control, and mechanical design for dexterous manipulation and human‑robot interaction. You will work on tactile sensing and manipulation policies for multi‑fingered hands, exploring reinforcement learning, imitation learning, and sim‑to‑real strategies on real hardware.

The role involves collaborating with hardware and controls teams, publishing in top venues, and mentoring engineers.

Qualifications

• Programming experience in Java, C++, or Python.
• Experience in tactile sensing, manipulation policy learning, or multi‑modal perception for robotics.
• Experience deploying learned policies on robotic hardware.

Responsabilités

• Design and train manipulation policies using tactile/force feedback and RL/IL or hybrids.
• Drive sim‑to‑real transfer for dexterous hands.
• Account for motor/actuator behavior in objectives and control strategies.
• Build simulation and on‑robot evaluation frameworks with clear benchmarks.

Connaissances

Java

C++

Python

Robotics research

Policy learning

Outils

MuJoCo

Isaac

URDF

Description du poste

Amazon is seeking talent to develop advanced robotics systems that combine AI, control, and mechanical design for dexterous manipulation and human‑robot interaction. You will work on tactile sensing and manipulation policies for multi‑fingered hands, exploring reinforcement learning, imitation learning, and sim‑to‑real strategies on real hardware.

The role involves collaborating with hardware and controls teams, publishing in top venues, and mentoring engineers.