Get roles like this, scored and matched, in your inbox.
Sign up free →Job Opening
Senior Staff Machine Learning Engineer, Perception, Simulation Realism
Waymo · Mountain View, CA · Posted 1 day ago
Set the technical bar for "Oracle" ground-truth perception across all AV software launches, scaling foundation-model signals to hundreds of new semantic categories over petabyte-scale data.
About the Team: The "Oracle" for Autonomous Driving
Our mission is to build the high-fidelity, platform-agnostic Ground Truth (GT) engine that powers the next generation of autonomous vehicle (AV) simulation and evaluation. We function as the "Oracle"—by reprocessing onboard logs offline using state-of-the-art ML Foundation Models, we generate semantic signals that significantly outperform real-time onboard software.
Since our inception in early 2025, we have moved from concept to a fully operational end-to-end ML solution. As we head into 2027, we are scaling this "Oracle" to become the backbone of all large-scale evaluation, enabling the business to test software at a level of rigor and realism previously thought impossible.
We are looking for a strategic technical leader to drive the ML perception architecture for Sim GroundTruth. This role sits at the critical intersection of cutting-edge perception ML research, foundation models, and massive-scale offline data engineering. As a Senior Staff Machine Learning Engineer, you will serve as the chief technical authority for our perception models, guiding and mentoring ML engineers to solve the "technical moat" of high-fidelity 3D perception and ML inference at a petabyte scale.
Key Responsibilities
…
This posting is sourced from Waymo's own career site — applying takes you there directly.
Get roles like this, scored and matched, in your inbox.
Sign up free →Related roles
This role is one of 567 Machine Learning Engineer openings and centers on Python, Numpy, Pytorch, and Jax. There are 381 open roles at Waymo right now.
Want roles like this found for you?
Set your Focus once. Career Engine screens every new opening every week.
A scored, explained shortlist — not a keyword match — delivered weekly. Free for one profile.