Intern Researcher – Embodied AI

About the role

Huawei Canada is seeking an Intern Embodied AI Researcher to conduct cutting‑edge research on safe and reliable embodied AI systems, focusing on robotic manipulation and autonomous decision making.

Key responsibilities

• Conduct cutting‑edge research on safe and reliable embodied AI systems, focusing on robotic manipulation and autonomous decision making.
• Develop novel reinforcement learning, imitation learning, and learning‑from‑demonstration algorithms to improve safety, robustness, and generalization of embodied agents.
• Design and evaluate methods that leverage negative data, failure cases, human feedback, and corrective demonstrations to improve policy learning.
• Develop and study Vision‑Language‑Action (VLA) and Vision‑Language Models (VLMs) for embodied intelligence and robot‑learning applications.
• Implement, train, and evaluate state‑of‑the‑art models in both simulation and real‑world robotic environments.
• Collaborate closely with researchers and engineers to translate research ideas into scalable systems and internal technology prototypes.
• Stay current with the latest advances in reinforcement learning, robot learning, embodied AI, foundation models, and AI safety, contributing technical insights to the research roadmap.
• Publish research findings at top‑tier machine learning and robotics conferences and contribute to high‑impact intellectual property.

Required profile

• Currently pursuing a Master’s or PhD in Computer Science, Robotics, Electrical Engineering, Machine Learning, Artificial Intelligence, or a related field.
• Proven research excellence with first‑author publications at top‑tier venues such as NeurIPS, ICML, ICLR, CoRL, RSS, CVPR, ICCV, ECCV, ICRA, or IROS.
• Strong background in reinforcement learning, including policy optimization, model‑based RL, offline RL, reward/value learning, or safe reinforcement learning.
• Hands‑on research experience with learning from demonstrations, imitation learning, inverse reinforcement learning, preference learning, or related paradigms.
• Experience developing, training, or evaluating Vision‑Language‑Action (VLA) models, Vision‑Language Models (VLMs), multimodal foundation models, or embodied AI systems.
• Strong programming skills in Python and deep learning frameworks such as PyTorch.
• Experience working with robot simulators, large‑scale training infrastructure, robotic platforms, or real‑world embodied AI systems.
• Demonstrated ability to independently drive research from idea formulation to implementation, experimentation, publication, and technology transfer.

Required skills

• Python
• PyTorch
• Reinforcement learning (policy optimization, model‑based, offline, safe RL)
• Imitation learning / learning‑from‑demonstration
• Vision‑Language‑Action (VLA) models
• Vision‑Language Models (VLMs)
• Multimodal foundation models
• Robot simulators
• Large‑scale training infrastructure
• Robotic platforms