Director of Engineering, Perception Data

The Perception Data team is the core function responsible for understanding vast amounts of Waymo driving data, curating them into datasets, and ensuring their availability and quality to accelerate Perception and development of the Waymo Driver. This includes areas such as efficient real-time onboard data selection, scalable data engine systems to power Perception's onboard models and foundation models, and large-scale data generation pipelines for Perception model development.

You will report to our Head of Perception.

You will:

• Keep close track and build understanding of evolving business and technology needs on data, including those that serve Waymo's product goals and foundation models
• Create data strategies that orchestrate the many facets involving data delivery: on-car real-time data sampling, offline data sampling, mining and sampling capabilities, data curation efficiency and quality, human labeling vs. autolabeling efficiency and quality tradeoffs, large-scale data generation systems, and storage
• Set up and evolve the working structure between multiple teams across Waymo, where we maximize reuse of infrastructure and tooling
• Responsible for ongoing delivery of diverse, fresh, and high-quality data for Perception's development, evaluation and model training
• Develop both data technologies and data tooling to support Waymo's increasing scale, and to support the diverse and advanced data requirements

You have:

• Track record in leading Data organizations that supported successful and large-scale products relying on ML, or
• Track record in leading a successful and large-scale quality-driven product that depends on the quality of data for its ML development
• Experience and mindset in working backwards from business goals to form the data strategies, while being extremely customer-focused - being a partner to the customer teams in shaping the best data plan for them together, and deliver
• Experience working leading horizontal programs
• Technical ability to dive deep into technical details, including complex systems, infrastructure, algorithms, and modeling techniques

We prefer:

• Experience leading ML-heavy quality-driven products, therefore having first-hand experience as users of data for ML development
• Familiarity with the data needs of computer vision models and VLMs
• Experience in data science to systematically understand the significance of data

#LI-Hybrid