Senior Software Engineer, Model Evaluation - Simulation

The Simulator team builds simulations of realistic environments crucial for testing and training the Waymo Driver. We use modern machine learning techniques to model the complexities of the real world, including the behavior of diverse agents (vehicles, pedestrians, cyclists, motorcyclists), intricate road networks, dynamic traffic control systems, and varying weather conditions. A core challenge is ensuring our simulations reflect reality. To guide the development of our simulator components, we create metrics and evaluation methodologies to measure realism. These metrics are important for assessing simulator quality, identifying critical sim-to-real gaps, and even serve as loss functions to train realistic simulation models – which, in turn, are fundamental to rigorously evaluating the Waymo Driver's performance and safety.

In this hybrid role you will report to an Engineering Manager.

You will:

• Develop novel methodologies and metrics to measure simulation realism across multiple simulator technologies (e.g., agent behavior, sensor rendering, environmental effects).
• Design scalable and efficient platforms and pipelines for measuring simulation realism across massive datasets.
• Research and implement systems that can automatically discover salient realism metrics to guide simulator development priorities.
• Collaborate with world-class engineering and research teams who develop and use large-scale ML models to enhance simulator fidelity.
• Play an important role in a collaborative engineering team dedicated to building a high-fidelity simulator that directly powers the safety validation and performance evaluation of the Waymo Driver.

You have:

• MS in Computer Science, Engineering, or a related technical field, or equivalent practical experience.
• 5+ years of industry experience in software development.
• Software Engineering Fundamentals:
  • Proficiency in programming in Python or C++, including experience with relevant libraries for data handling and computation (e.g., NumPy, Pandas for Python).
  • Experience with SQL for data querying and analysis.
  • Experience with software design principles, coding best practices, testing methodologies, and version control software.
• Quantitative Application & Metric Development:
  • Identify potential system performance or realism gaps (e.g., from specific examples or initial data exploration) and create data-driven approaches to investigate them.
  • Experience designing, implementing, and interpreting quantitative analyses or metrics to systematically evaluate system behavior and validate hypotheses (using concepts such as distributions, confidence intervals, etc.).
  • Demonstrated experience taking quantitative findings through to implementation, resulting in productionized metrics, automated evaluation tools, or improvements to validation systems.
• Software Development for Data/Evaluation Pipelines:
  • Experience building software pipelines for data processing, system evaluation, or metric computation, in the context of large-scale systems.

We prefer:

• Demonstrated expertise and influence in designing quantitative evaluation methodologies and building scalable data processing/evaluation infrastructure.
• Experience navigating complex technical and product landscapes, defining technical strategy, and creating roadmaps.
• Experience with simulation systems, robotics, or autonomous vehicles.

#LI-Hybrid