Senior Machine Learning Engineer, Runtime & Optimization

The ML Platform team at Waymo provides a set of tools and technologies to support and automate the lifecycle of the machine learning workflow, including feature and experiment management, model development, debugging & evaluation, optimization, deployment, and monitoring. These efforts have resulted in making machine learning more accessible to teams at Waymo, including Perception, Behavior Prediction, Planner, Maps and Research, ensuring greater degrees of consistency and repeatability, and addressing the "last mile" of getting models into production and managing them once they are in place. We work hand in hand with machine learning experts in all parts of the company as well as our collaborators across Alphabet.

We are looking for engineers with ML software or ML systems expertise to help us improve compute performance on our car. You'll work across the entire ML framework/compiler stack (e.g. JAX, XLA, Triton, and CUDA), as well as system-efficient deep learning models. You will be pleasantly challenged with deploying a hugely diverse set of Waymo ML models on limited computation resources. Non-exhaustive examples of the types of work you will work on:

• Collaborate with ML practitioners on models for perception, behavior prediction and planner, to understand their models and modify model architectures to run faster on the car.
• Deep dive into the whole stack of ML software stack, from custom ops, framework/ ML compiler, to low level libraries. Analyze numeric behaviors. Make sure inference results are stable and consistent across multi-devices. Develop tools/system software for optimal resource usage, hardware efficiency, and platform reliability in a ML serving system.
• Analyze the ML workload performance; Apply model optimization and efficient deep learning techniques to models; Develop highly optimized ML operator libraries.
• Build tools to benchmark, optimize and productize deep learning models for a streamlined and robust onboard and offboard deployment.

At a minimum, we'd like you to have:

• B.Sc in Computer Science, Mathematics or a related field
• 3+ years of industry experience
• Strong C++ programming skills
• Experience in developing on or using deep learning frameworks (e.g., PyTorch, TensorFlow, ONNX, etc.)
• Passion for developing and optimizing ML software stacks for modern ML accelerator architectures (framework, runtime library, ML compiler, efficient deep learning etc.)
• Working knowledge on system performance, GPU optimization or ML compiler

It's preferred if you have one of the following:

• M.Sc or PhD in Computer Science, Mathematics or a related field.
• Strong Python programming skills.
• Strong experience with efficient deep learning and/or model optimization techniques such as quantization, NAS, distillation.
• Solid experience with designing, training and debugging deep learning models to achieve the highest scores/accuracies.
• In-depth knowledge of ML frameworks, ML compiler and IRs (Triton, HLO, MLIR, Relay) or modern ML system architectures.

#LI-Hybrid