Role description
You will co-develop the AI Copilot software that allows users to control humanoid robots through natural language. Your work will focus on motion generation, learning-based control, and AI systems that translate LLM instructions into robot motion and actions. You will develop algorithms and software that enable humanoid robots to reliably perform complex tasks in dynamic real-world environments.
Your work will involve designing reinforcement-learning and imitation-learning pipelines, developing simulation workflows in MuJoCo and Isaac Lab, and deploying trained policies to physical robots. You will address challenges such as sim-to-real transfer, policy robustness, and policy performance across diverse real-world tasks and environments.
This position is in an early-stage startup environment. You should be ready for ambiguity, tight schedules, shifting priorities, rapid iteration, and taking responsibility for executive-level decisions that shape the final product-market fit.
Responsibilities
- Develop AI-based control and motion generation algorithms for humanoid robots.
- Design and optimize reinforcement learning and imitation learning pipelines.
- Build and maintain simulation workflows using MuJoCo, Isaac Lab, or similar physics engines.
- Deploy and validate trained policies on physical robots.
- Improve sim-to-real transfer, policy robustness, and task performance across real-world environments.
- Develop robotics software using Python, Linux, Docker, and GPU-based compute infrastructure.
- Create and maintain robot models, simulation assets, and training environments.
- Collaborate with the engineering team to integrate new capabilities into the AI Copilot platform.
Main requirements
- 2+ years of experience in robotics, simulation, or motion planning.
- Strong Python programming skills.
- Experience with robot simulation environments such as MuJoCo, Isaac Lab, Isaac Sim, or similar.
- Experience with reinforcement learning (PPO, SAC) or imitation learning.
- Solid understanding of robot kinematics, dynamics, and control.
- Experience working with Linux and NVIDIA GPU-based computing.
Other requirements
- Experience developing robotics or machine learning software.
- Experience with Docker and containerized development environments.
- Fluent English.
Nice to have
- Experience with humanoid robots, legged robots, or robotic manipulation.
- Familiarity with NVIDIA Isaac Lab, GR00T, Newton, or similar AI frameworks.
- Experience with ROS2.
- Experience creating robot models using URDF or MJCF.
- Published research in robotics or machine learning.
- Full-stack development experience for web or APIs (React, FastAPI).
