Senior Machine Learning Engineer, Runtime and Serving

We are looking for engineers with ML software & systems expertise to help b uild the next generation Waymo onboard ML inference engine for Waymo fundamental model. You'll work across the entire ML stack from the system perspective, from efficient deep learning models, model compression, ML software (e.g. JAX, XLA, Triton, and CUDA), to .

You will be pleasantly challenged with deploying Waymo ML models on limited computation resources. In this hybrid role, you will report to the Senior Manager of Runtime and Optimization.

- Architect and develop an efficient, high-performance ML runtime and serving system tailored for both onboard autonomous vehicle compute and large-scale, offboard data center environments.

- Lead the integration and feature development for ML inference runtimes across both domains, balancing the strict real-time latency and memory constraints of onboard systems with the high-throughput, highly concurrent demands of offboard serving fleets.

- Drive the strategic migration of ML workloads toward a JAX-native runtime architecture, which includes extending and modifying underlying ML compilers and runtimes (e.g., OpenXLA/PjRT, TensorRT).

- Collaborate with world-class Waymo ML practitioners across perception, planner, and research to analyze system-level ML workloads and apply hardware-aware compute optimizations.

- Design and build robust tooling for profiling, benchmarking, and identifying system-level bottlenecks across the end-to-end ML software stack.

- 5+ years of professional software engineering experience focused on building, scaling, or maintaining ML systems and infrastructure.

- 3+ years of production experience in Python and major deep learning frameworks (e.g., PyTorch, JAX).

- Experience optimizing ML software for hardware accelerators (e.g., GPUs, TPUs, custom silicon).

- Experience building low-latency, highly concurrent distributed backend systems.

- Experience modifying ML compilers, runtimes, or inference engines (e.g., TensorRT, ONNX Runtime, OpenXLA/PjRT, TVM).

- Experience building or scaling LLM serving systems, including expertise in distributed inference and performance optimization (e.g., KV/prefix caching, continuous batching).

- Experience with custom kernel development (e.g., CUDA/CUDA Tile, Triton, JAX/Pallas).

- Experience architecting unified serving APIs and optimizing tensor buffer management (e.g., zero-copy data transfer, shared memory) for complex, multi-model inference pipelines.