2 months ago
Base Salary
$251k - $310k/yr
Responsibilities
- Conduct experiments to train and deploy multimodal large language models and world models for 3D perception using camera, LiDAR, and radar data.
- Partner with Waymo engineering and research teams to deploy models and implement efficient continuous-training workflows using newly collected data.
- Develop and apply quantization, pruning, knowledge distillation, and efficient attention techniques.
- Develop and maintain scalable training and evaluation data pipelines processing multiple data sources.
- Design and implement evaluation frameworks for perception models.
- Build infrastructure for large-scale model distillation and teacher-model bulk inference.
- Experiment with model partitioning and sharding strategies to improve scalability and efficiency.
- Build and maintain performance-analysis, profiling, and debugging tools for ML models.
Requirements
- Master’s or PhD in Computer Science, Machine Learning, Robotics, or a similar technical field.
- At least 4 years of industry or postdoctoral research experience in reinforcement learning or foundation models.
- Proficiency implementing scalable, distributed, and performant model-training flows, including data parallelism and FSDP.
- Proficiency in JAX and Flax, with potential experience in TensorFlow or PyTorch.
- Willingness to work with globally distributed inference infrastructure.
- Hands-on experience optimizing training and inference for Transformer architectures.
- Preferred: PhD focused on reinforcement learning, foundation models, or multimodal learning.
- Preferred: substantial contributions to high-impact industry AI projects.
- Preferred: experience with generative models for world models, images, videos, or 3D, including diffusion or autoregressive models.
- Preferred: contributions to frameworks and libraries that improve training speed and scalability, such as JAX, Gemax, and XManager.
Benefits
- Hybrid work schedule.
- Health, dental, vision, life, and disability insurance.
- 401(k) with company match.
- 20 vacation days per year for the first five years.
- Sick time of 40 hours per year, where applicable.
- 28–30 weeks of maternity leave.
- 18 weeks of baby bonding leave.
- 13 paid holidays per year.
- Eligibility for annual bonus, equity incentive plan, and company benefits program.
Tech Stack
PyTorchTensorFlow
Categories
About Waymo
Waymo develops the Waymo Driver, a full-stack autonomous driving system, and operates the Waymo One robotaxi service that offers paid, driverless rides in select U.S. cities. The company monetizes through ride-hailing fares and partnerships integrating its technology on multiple vehicle platforms. Founded in 2009 as Google’s self-driving project, Waymo is headquartered in Mountain View and is an Alphabet subsidiary, with over 10 million rider-only trips and 100+ million autonomous miles on public roads.