AMI Scientist/Interns - Geometry and 3D Vision
Core
Researching world model-based AI systems that understand, predict, and plan in the real world using 3D geometry and video data.
Role type
Research Scientist (3D Vision and World Models)
Builds
Frontier world model-based AI systems
Domain
Artificial Intelligence, Computer Vision, Robotics
Deliverable
production ML models
Required skills
Python, machine learning fundamentals, large-scale training, accelerator-based compute (GPU/TPU), experimental design and analysis
Preferred skills
3D/4D reconstruction, SLAM, SfM, depth and pose estimation, point tracking, rendering/simulation engines (Blender, Isaac Sim), deep learning frameworks (PyTorch, JAX), open-source project development
Technologies
Python, PyTorch, JAX, Blender, Isaac Sim
Responsibilities
Develop self-supervised learning methods for video and high-dimensional signals; Design new architectures for predicting world dynamics; Build scalable algorithms for video data pre-processing and curation; Create evaluation frameworks for benchmarking world model understanding and planning; Develop efficient algorithms for model-based planning and reasoning
Seniority
Individual Contributor (Researcher)
