Robotics Software Intern
You will collect and curate teleoperation data, fine-tune VLA models, and run evaluations in both simulation and on physical robots. Additionally, you will support physical hardware setups including cameras, sensors, and control loops.
- Remote
- Canada
- Posted Jul 21, 2026
- 1 position
Job summary
Requirements: USA-based. Do not apply if you're not in the USA Experience with version control, preferably GitHub Solid Python; familiarity with PyTorch (or a strong ML coursework background) A creative and innovative problem solver Python, PyTorch, Linux, Git; bonus for C++, ROS, or CUDA You'll help collect and curate teleoperation data, fine-tune and train VLA models (using frameworks like LeRobot, OpenVLA, or π0), run evaluations in simulation and on real robots, and support the physical setup — cameras, sensors, calibration, and the control loop that turns model outputs into motion. This is an unpaid position. Extra: Experience with imitation learning, reinforcement learning, or transformer-based models Familiarity with a robotics simulator (Isaac Sim, MuJoCo, or LeRobot sim environments) Working with Claude Code and vibe-coding platforms or third parties Collaborate closely with the team to evaluate and improve the accuracy and reliability of the robot's perception and manipulation Any prior robotics project — a personal arm, drone, ROS work, or a robotics class
What you’ll do
You will collect and curate teleoperation data, fine-tune VLA models, and run evaluations in both simulation and on physical robots. Additionally, you will support physical hardware setups including cameras, sensors, and control loops.
Requirements
Candidates must be based in the USA and possess strong Python skills along with experience in version control and machine learning. Familiarity with robotics frameworks, simulators, or prior robotics projects is highly preferred.
Other relevant skills
Identified from the job description. Confirm important requirements above.
- Python
- PyTorch
- Linux
- Git
- C++
- ROS
- CUDA
- Imitation Learning
- Reinforcement Learning
- Transformer-based Models
- Isaac Sim
- MuJoCo
- LeRobot
- VLA Models
- Version Control
- Machine Learning
Job areas
- Software
- Engineering
- Technology
- Science & Research
- Data & Analytics
Additional details
- Minimum experience
- 0+ years
- Posting language
- English
- Working hours
- 40 hours per week
- Location requirements
- Country, United States
- Application method
- Direct apply is available