Waymo is an autonomous driving technology company with a mission to make it safe and easy for people and things to get where they’re going. Since our start as the Google Self-Driving Car Project in 2009, Waymo has been focused on building the Waymo Driver—The World’s Most Experienced Driver™—to improve everyone's access to mobility while saving thousands of lives now lost to traffic crashes. Our Waymo Driver powers Waymo One, our fully autonomous ride-hailing service, as well as Waymo Via, our trucking and local delivery service. To date, Waymo has driven over 20 million miles autonomously on public roads across 25 U.S. cities and conducted over 20 billion miles of simulation testing.
The Multiverse team is responsible for evaluating the safety of new releases before putting new version on the road, especially the one without safety driver. This is a very complex process running at an enormous scale with unique challenges (we can't just run A/B experiments) that has been built from years of experience. We are in the process of scaling this process to be able to tackle multiple cities and platforms at once.
We have multiple challenging roles with either:
- A focus on quality: How do we evaluate driving in Simulation? How do we find the very long tail of events? How do we deal with the messy real world? This would be a great role for a person with experience in search quality or ads quality or evaluation. Similarly, experience in machine learning (including deep learning, reinforcement learning, etc), data science, statistics, are a bonus but not required.
- A focus on Infrastructure: We run some of the largest jobs, distributed across 100+ Data Centers, and need to handle CPU/TPU interfaces without dropping one of the critical rare events. We're at the point where we need to rebuild significant part of the system while in-flight. This would be a great fit from someone with experience in search/ads backend for example.
Some of the projects the team is tackling include:
Designing and integrating cutting edge research in large scale production systems (ML prediction on TPU)
Traffic Situational Modeling for making judgements about the safety and quality of driving (leveraging geometric / kinematics-based, heuristic and ML learned techniques)
Extracting fast, leading indicators both for use in direct optimization and evaluation of our vehicle's behaviors
Leveraging millions of miles of driving logs, balancing deep and broad evaluation strategies and signals for assessment
Improving and evaluating agent behavior for realistic simulations
We are looking for quality-minded engineers to work with an experienced team. The following types of experience are relevant to roles on this team.:
Primary Qualifications (at least one of)
Experience in the area of quality, experimentation, evaluation, statistical gotchas, machine learning, etc.
Passionate about eval work: building internal tools, scalable systems, and data processing pipelines. Making our system better through data: "If you can't measure it, you can't improve it!"
Experience with large scale system design, e.g. integration of ML research in production systems
MS or PhD in a quantitative field
Experience with Machine Learning (especially with TensorFlow) modeling and / or infrastructure
Interest and/or experience with data labeling, or machine learning data
Note: The following is only applicable for roles hired in the US. As we keep the safety of our employees and our communities top of mind, COVID-19 vaccinations continue to be a critical prevention measure to help end the pandemic. Because of this, Waymo requires all US-based employees who perform work onsite for Waymo to be fully vaccinated against COVID-19. If you are unable to be vaccinated, Waymo will provide a reasonable accommodation consistent with applicable law.