NVIDIA Isaac develops tools to bring accelerated computing to robotics scientists and developers. We are looking for a self-motivated, hard-working, and innovative internship candidate to assist us in developing novel approaches to solve one of the hardest problems in robotics: spatial intelligence on humanoid robots. In particular, this internship will examine how recent advances in real-time scene understanding, neural, and optimization-based reconstruction can be applied to robotic manipulation. The anticipated results of this internship will be technical contributions to real-time GPU-accelerated spatial intelligence tools deployed on robot manipulators.

You will join NVIDIA’s Robotics Software team in Zurich, Switzerland, and work on real-time scene understanding and 3D reconstruction for robots. You will collaborate with the Isaac 3D perception team to develop GPU-accelerated tools and deploy them in simulation and on real robot hardware. The role requires on-site presence in our Zurich office.

‍

‍

What you'll be doing:

• Review literature and investigate state-of-the-art methods for real-time neural, optimization-based, and feed-forward reconstruction that could solve the unique challenges of robotic manipulation.
• Collaborate with the team on implementing novel reconstruction techniques with application to robotic manipulation.
• Work in simulation, where ground-truth data is available to measure the system’s performance.
• Repeatedly quantify and improve the performance of the system.
• Deploy algorithms to real robot hardware.
• Work in a highly collaborative environment with a small robotics team.

‍

‍

What we need to see:

• Experience in scene understanding, 3D reconstruction, or robot mapping.
• Currently pursuing a degree in Computer Science, Robotics, Engineering, or a related field (Ms or PhD).
• High proficiency in at least one programming language (c++ or python).
• Knowledge of common tools and libraries for computer vision (i.e. pytorch, opencv).
• Academic classes or coursework in machine learning, computer vision and/or robotics systems.

‍

‍

Ways to stand out from the crowd:

• Experience in neural reconstruction.
• Previous internship experience with robotics and/or computer vision systems.
• Previous contributions to a related software project.
• Academic publications in a relevant domain.
• Experience with tools used in the NVIDA robotic’s ecosystem (CUDA, warp, Isaac Lab, Isaac Gym, ROS).

‍

Applications will be accepted until: December 11, 2025.

‍