AI Scientist – World Modeling

About the role

AMI is advancing AI agents that can understand, predict, and plan using an internal world model. As an AI Scientist you will push the state‑of‑the‑art in world‑model research, focusing on video and multimodal data to enable agents to act safely in complex environments.

Key responsibilities

• Design and implement self‑supervised learning methods for continuous, high‑dimensional signals such as video.
• Develop new neural architectures that efficiently predict world dynamics.
• Build scalable pipelines for preprocessing and curating large video datasets.
• Create evaluation frameworks to benchmark model understanding, prediction, and planning.
• Research and prototype efficient algorithms for model‑based planning and reasoning.

Required profile

• Bachelor’s degree or equivalent experience in Computer Science or a related field.
• Proficiency in Python and ability to run independent experiments.
• Solid understanding of machine‑learning fundamentals, large‑scale training, and GPU/TPU compute environments.

Required skills

• Python programming.
• Deep‑learning frameworks such as PyTorch or JAX.
• Experience with self‑supervised learning, video or multimodal model architectures, and planning algorithms (preferred).