Senior Software Engineer, Waymo

• M.S. in Computer Science, Machine Learning, Robotics, Statistics, Physics, Math or another quantitative area
• 5+ years of industry experience building and evaluating large-scale ML systems
• Proficiency in Python
• Familiarity with one of the modern deep learning frameworks (e.g. JAX, Tensorflow)
• Experience with tools for large-scale parallel compute and data manipulation (Apache Beam, SQL)

Desirable

• The Planner Evaluation team works on one of the key challenges in autonomous driving: improving the quality of the software that drives the car
• We are looking for quantitative-minded software engineers to help us understand and improve the performance of our ML and robotics systems
• If you are excited about applying a large volume of extremely rich data to improve core technology at the very heart of autonomous driving, this is the place for you
• In this Software Engineering role, you will report to the Manager of Planner Evaluation
• Develop evaluation methodologies for Foundation Models and production models: Research and propose new ways to assess the quality, safety, and realism of ML models used to generate driving plans and agent trajectories
• This includes exploring new metrics, statistical analysis, and innovative validation techniques applicable to both Waymo Research's cutting-edge Foundation Models and the production models deployed in the Waymo Driver
• Contribute to the design and development of robust ML evaluation tools: Design, build, and improve evaluation tools that enable model hillclimbing and inform production model release decisions
• This involves collaborating with engineers to define requirements, designing new features, and providing support for these tools, ensuring they meet the needs of both research and production environments
• Build evaluation products that advance business needs: Partner with teams in Planner and Research to launch new and improved ML evaluation products that directly address business requirements, ensuring that evaluations are aligned with the overall goals of the Waymo Driver and company objectives
• Leverage Waymo's industry-leading datasets and evaluation frameworks: Utilize extensive datasets and established evaluation frameworks to conduct robust and reliable evaluations, ensuring that model improvements translate into measurable gains in fleet-wide performance
• Employ creative simulation strategies to measure the driving performance of generative AI models: Identify potential edge cases, and provide reliable performance insights that inform model development and deployment
• Build data pipelines for signal discovery, data labeling, feature extraction and metric computation based on large-scale simulations