Machine Learning: World Models
Build world models that predict the consequences of robot actions and enable long-horizon planning.
What you'll do
- Develop generative world models for physical environments
- Design architectures for action-conditioned prediction
- Create training pipelines that leverage real-world deployment data
- Research methods for compositional and transferable representations
- Evaluate world model accuracy across diverse physical scenarios
What we're looking for
- Experience with generative models, video prediction, or dynamics modeling
- Strong background in deep learning and probabilistic inference
- Proficiency in PyTorch and large-scale model training
- Publications in top ML/AI/Robotics conferences preferred
- PhD in Machine Learning, Computer Science, or related field preferred