Research Scientist, World Model Post-Training
About this role
Conducting applied world model research and development forms the core of daily activities in this hybrid position on the AI Foundations team at Waymo. Team members focus on post-training generative models to solve complex challenges in autonomous driving and support safe vehicle operations across diverse conditions. Practitioners design experiments, train large models, and share findings with peers and leadership while leveraging artificial intelligence to automate workflows and write documentation.
Waymo operates as an autonomous driving technology company that develops the Waymo Driver to improve mobility and traffic safety. Individuals joining the AI Foundations team collaborate closely with researchers across Alphabet on advanced machine learning solutions. Ideal candidates bring advanced academic credentials in deep learning along with hands-on experience in modern frameworks and generative architectures.
What you will do
- Conduct applied research and development focused on world models for autonomous driving.
- Post-train foundational world models using large-scale generative techniques.
- Design and execute experiments to train and evaluate large generative models.
- Present research findings and project results to peers and leadership teams.
- Utilize artificial intelligence tools to write high-quality code, unit tests, and documentation, as well as automate workflows.
What you need
- Hold a Master's degree or PhD specializing in deep learning with a focus on generative or world modeling.
- Bring a minimum of three years of experience working with modern deep learning frameworks such as JAX, TensorFlow 2, or PyTorch.
- Demonstrate hands-on experience with diffusion or flow matching models.
- Possess practical background in reinforcement learning.
- Show proven experience training large foundational models.
Nice to have
- Experience with post-training of foundational models
- Experience with reinforcement learning for diffusion or flow matching models
- Experience with distillation
- Experience with autonomous vehicles
Benefits
- Discretionary annual bonus program
- Equity incentive plan
- Company benefits program
Skills and tech stack
Questions candidates ask
What is the work policy for this role?
This is a hybrid role based across US locations including Mountain View, CA, San Francisco, CA, and New York, NY.
What is the salary range for the position?
The expected base salary range for this full-time position across US locations is $213,000 to $263,000 USD, with actual starting pay based on job-related factors.
What frameworks and models are used on the team?
The team works with modern deep learning frameworks such as JAX, TensorFlow 2, and PyTorch, along with diffusion, flow matching, and large foundational models.
Who does this position report to?
In this hybrid role, you will report to a Software Engineering Manager.