Simulation (Isaac Sim) Intern
Pune, IndiaInternship
We are looking for a hands-on Simulation Intern who can help build, configure, and test NVIDIA Isaac Sim environments for our robot programs, working alongside our senior engineers.
This internship is suited for a student or early-career engineer who is curious about simulation, comfortable in Python and Linux, and wants to see simulated scenes, sensors, and synthetic data used to develop and validate real robots.
What you'll work on
- Building and editing USD scenes, robot assets, and environments in NVIDIA Isaac Sim.
- Importing and configuring robots from URDF or MJCF, including joints, physics, and collisions.
- Setting up simulated sensors such as RGB and depth cameras, LiDARs, and IMUs.
- Connecting simulations to robot software through the Isaac Sim ROS 2 bridge.
- Generating synthetic datasets with Replicator, including labels and annotations.
- Domain randomization of lighting, textures, object poses, and camera parameters.
- Supporting robot learning experiments in Isaac Lab where they fit the program.
- Testing scenes, comparing simulated behaviour against real robots, and documenting results.
You will work with senior engineers to turn a test requirement into a working scene, check that it behaves as expected, and report the gaps between simulation and the real robot.
What we're looking for
- Working knowledge of Python.
- Comfort with Linux, the terminal, and Git.
- Genuine curiosity about robot simulation, physics, and 3D scenes.
- Basic understanding of coordinate frames, transforms, and robot joints.
- Willingness to read documentation and sample code to learn new tools independently.
- At least one project in simulation, robotics, graphics, or 3D that you can explain in detail.
- Clear written communication when reporting progress and issues.
Good to have
- Simulation platforms: NVIDIA Isaac Sim, Isaac Lab, Gazebo, MuJoCo.
- Scene and asset tools: OpenUSD, Omniverse, Blender, URDF / MJCF.
- Synthetic data: Omniverse Replicator, domain randomization, COCO / KITTI-style labels.
- Robot software: ROS 2, RViz, MoveIt 2.
Experience with reinforcement learning, sim-to-real transfer, GPU workstations, or computer vision datasets is an added advantage.
Who will fit this role
You should be comfortable taking a simulation task from a short description to a working, tested scene, and explaining clearly how closely it matches the real robot.
We are looking for someone who enjoys building virtual worlds that are useful to real robots, and who cares whether the simulation is actually right.
This is an internship position.
If this is your seat
Send the application. We read every one, and we write back when it's a fit.