Research Engineer/Scientist, World Models
We are building a general AI for robots that learns in the real world with minimal human supervision. Our robots collect massive amounts of data autonomously containing valuable failures and recoveries while minimizing unscalable human language advice and teleoperation. We develop world-model based reinforcement learning algorithms to learn reliable policies from this dataset and improve beyond human performance.
You might thrive in this role if you love being on the cutting edge of world model research and can iterate fast on new algorithm designs.
Responsibilities
- Develop and implement new algorithms for training world models
- Perform training of deep neural networks on large GPU clusters
- Collaborate with robot learning researchers to train robot foundation models
Preferred Qualifications
- Deep technical knowledge and research experience in large generative models
- Experience implementing learning algorithms for video or image generation
- Extensive experience in Python and at least one deep learning library such as PyTorch, JAX, TensorFlow etc.
- Publications in top conferences (e.g. NeurIPS, ICLR, ICML, CVPR, RSS, CoRL, ICRA)