jobapplierResearch Scientist (World Model Post-Training)
- New York, United States
- Mid Level
- 3+ years
- PhD or doctorate
- Posted
Skills
Required
- reinforcement learning
- diffusion / flow matching models
- generative / world modeling
- modern deep learning frameworks
- JAX
- TensorFlow
- Pytorch
Preferred
- autonomous vehicles
- Leverage AI
- unit tests
- documentation
- automate workflows
- generative modeling
- robust evaluation
About Waymo
Join Waymo, a leader in autonomous driving technology. As a Research Scientist, you will conduct applied world model research and development, post-train Waymo's world model, design experiments, and present results to peers and leadership. You will also leverage AI to write high-quality code, unit tests, and documentation, and automate workflows. Enjoy a hybrid work model, competitive compensation, and a comprehensive benefits package.
Conduct applied world model research and development, including post-training of Waymo’s world model and designing experiments.
Responsibilities
- Leverage AI to write high-quality code, unit tests, and documentation, and automate workflows.
- Collaborate with other research teams in Alphabet, focusing on areas such as reinforcement learning, generative modeling, and robust evaluation.
Requirements
- Masters or PhD in deep learning in generative / world modeling
- Experience training large foundational models
- Experience with diffusion / flow matching models
- 3+ years of experience with modern deep learning frameworks (e.g., JAX, TensorFlow 2, Pytorch)
- Experience with reinforcement learning
- Experience with post-training of foundational models
Preferred
- Experience with reinforcement learning for diffusion / flow matching models
- Experience with distillation
- Experience with autonomous vehicles
Pay
USD 213,000 - 263,000