Scientist/Intern – Geometry and 3D Vision

About the role

AMI is building AI agents that understand and plan in the real world using internal world models. We are looking for scientists and interns to advance world‑model research, focusing on 3D geometry extracted from video and other high‑dimensional sensor streams.

Key responsibilities

• Develop self‑supervised learning methods to efficiently learn from video and continuous sensor data.
• Design new architectures that predict world dynamics from high‑dimensional inputs.
• Create scalable pipelines for preprocessing and curating large video datasets.
• Build evaluation frameworks to benchmark understanding, prediction, and planning capabilities of world models.
• Implement efficient algorithms for model‑based planning and reasoning in dynamic environments.

Required profile

• Bachelor’s degree or equivalent experience in Computer Science or a related field.
• Proficiency in Python and ability to design, run, and analyse experiments independently.
• Solid understanding of machine‑learning fundamentals, large‑scale training, and GPU/TPU compute environments.
• Deep expertise in at least one area such as 3D/4D reconstruction, SLAM, SfM, depth or pose estimation, or point tracking.
• Hands‑on experience with a rendering or simulation engine (e.g., Blender, Isaac Sim) and scripted scene generation.
• Track record of research contributions through publications, model releases, or open‑source projects.

Required skills

• Python
• PyTorch
• JAX
• GPU
• TPU
• Blender
• Isaac Sim

What we offer

• Opportunity to work on cutting‑edge AI research in a global, multidisciplinary team.
• Access to state‑of‑the‑art compute resources and large‑scale video datasets.
• Collaborative environment that bridges scientific rigor with engineering excellence.