Humanoid robotics and physical AI engineer

RAISEBOX is a Physical AI R&D company focused on training humanoid robots for real-world industrial work. The team develops advanced dexterous manipulation, navigation, and whole-body control skills for the Unitree G1 humanoid robot. These capabilities are learned from human teleoperation demonstrations and high-fidelity simulation environments such as MuJoCo, NVIDIA Isaac Lab, and GR00T, and then transferred to the physical robot.

• designing, implementing, and testing policies for humanoid robot control, perception, and learning, with a focus on dexterous manipulation and navigation in industrial environments.
• developing and integrating machine learning models, including neural networks, with robotic hardware and simulation platforms
• building robust software components and tools to support teleoperation, data collection, and sim-to-real transfer.
• debug and optimize code, run experiments in simulation and on real robots, analyze performance metrics, and document results to inform iterative improvements.
• Participation in code reviews, technical discussions, and joint problem-solving sessions will be key
• maintaining safe and reliable operation of humanoid systems in the lab.

• Strong foundation in Computer Science and related fields, with experience applying Pattern Recognition techniques to real-world problems.
• Hands-on experience with Neural Networks and modern machine learning frameworks for control, perception, or reinforcement learning in robotics.
• Proficiency in Software Development, including clean coding practices, version control (e.g., Git), and debugging in languages such as Python and C++.
• Background in robotics, control systems, or physical AI, preferably with humanoid or legged robots.
• Ability to work on-site in Lisbon, collaborate in a research-oriented team, and communicate clearly in English; Portuguese is a plus but not required.
• Bachelor’s, Master’s, or PhD in Computer Science, Robotics, Electrical/Mechanical Engineering, or a related technical discipline, or equivalent practical experience.
• Будет плюсом: Familiarity with Natural Language Processing (NLP) and its integration with robotic or interactive systems is a plus; Experience with robotic simulation tools (e.g., MuJoCo, Isaac Lab) and sim-to-real transfer methods is highly beneficial

• Formal employment contract;
• In-house lab with humanoid robot, NVIDIA DGX Spark cluster and access to the Taguspark and IST ecosystem; real technical autonomy in a startup company;
• 30.000 EUR gross, possible equity participation.