Staff ML Engineer, Generative Model Performance & Efficiency
About The Position
Waymo’s Simulator Team develops realistic simulations for testing, training, and validating the Waymo Driver. The team uses advanced generative and reconstructive machine learning to model real-world environments, including agents, roads, traffic systems, weather, and vehicle sensors. This role focuses on improving the fidelity, scalability, controllability, and richness of simulations through efficient generative models and 3D world modeling. The position reports to a Senior Staff Engineering Manager. The engineer will analyze and optimize model architectures and training and inference performance; apply compression and efficiency techniques; optimize code for TPU and GPU accelerators; develop scalable model partitioning strategies; build low-latency serving systems and efficient training pipelines; and create tools for profiling and debugging machine-learning workloads.
Requirements
- MS or PhD in Computer Science, Machine Learning, Robotics, or a related field.
- At least 5 years of experience with deep learning architectures, especially Transformers, Diffusion Models, and MoEs, as well as algorithms and optimization techniques.
- Proficiency in JAX and Flax, and potentially TensorFlow or PyTorch.
- Expertise using profiling tools such as XProf, Perfetto, and NVIDIA Nsight to diagnose performance issues in machine-learning workloads.
- Hands-on experience with quantization, pruning, distillation, and other model-compression methods.
- Strong Python programming skills and potentially C++, with experience in software development best practices.
Nice To Haves
- Knowledge of TPU and GPU architectures and how to optimize code for them.
- Familiarity with ML compilers such as XLA and understanding of how they translate high-level code into efficient hardware instructions.
- Understanding of training and serving models across multiple devices and machines.
- Experience contributing to frameworks and libraries that improve training speed and scalability, such as JAX, Gemax, or XManager.
Responsibilities
- Analyze model architectures and identify training and inference bottlenecks, including memory bandwidth, compute, and communication.
- Apply and develop quantization, pruning, knowledge distillation, and efficient attention techniques.
- Optimize model code for hardware accelerators such as TPUs and GPUs, using compiler features and low-level libraries such as XLA.
- Experiment with data, tensor, pipeline, and expert parallelism and other model partitioning and sharding strategies to improve scalability and efficiency.
- Design and implement low-latency, high-throughput serving solutions for generative models and optimize training pipelines to reduce training time.
- Build and maintain tools for performance analysis, profiling, and debugging of machine-learning models.
Benefits
- Discretionary annual bonus program
- Equity incentive plan
- Company benefits program
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
