Our Aurora Voices series celebrates the people and teams whose unique experiences, backgrounds, and voices bring Aurora’s mission to life.
Meet Sophie, VP of Simulation, Test, and Validation at Aurora
At Aurora, we use simulations to create virtual models of the world to test how the Aurora Driver reacts in various situations. These virtual tests help us identify and remediate potential issues and challenges during development, helping our vehicles to perform safely and reliably when driving on the road.
As VP of Software Engineering, Sophie leads Aurora’s Simulation, Test and Validation team—a 160-person organization that conducts millions of simulation tests each day. She also founded Aurora’s first data science community, promoting a strong data-driven culture in the company.
Sophie joined the autonomous vehicles industry after 18 years at Google, pursuing an opportunity to make an impact in an industry that will transform the world. Read on to learn more about her story and how her personal passions on the golf course help inform her everyday decisions.
Tell us about your role at Aurora.
I lead Aurora’s Simulation, Test and Validation team in the Software Engineering organization. Our mission is to 1) develop and advance the use of world-class simulation technology, 2) deliver the workflow, tooling, metrics, and analysis our engineering teams need for testing, triage, and evaluation, and 3) accelerate autonomy development and validation of the Aurora Driver.
One of my early accomplishments at Aurora was bringing together the data science and data analytics teams to create Aurora’s first data science community. The Data Science@Aurora community mission is to promote a strong data-driven culture across all teams, unlock value from data, and improve decision-making at all levels with data insights.
At Google, I saw first-hand the impact of a strong data science community and knew I wanted to create that same meaningful experience within Aurora. Our community is made up of Aurora’s data science leads and teams from across product, safety, software, and operations, including engineers and researchers who are interested in or working on machine learning or data analysis. We share best practices, influence decisions in Aurora’s product strategy and development, and drive greater efficiency across business, engineering, and operations.
How do we use simulation and data science at Aurora?
Simulation is at the core of our development and enables us to make fast progress safely. Aurora generates hundreds of thousands of virtual tests (or simulation scenarios) that allow us to understand how the Aurora Driver may perform in different situations, including those we rarely experience while driving on road, such as collisions. These scenarios are specifically designed to help us find edge cases and catch errors in our software before putting our vehicles on the road, helping the Aurora Driver learn to behave safely and responsively while navigating complex scenarios.
Running millions of simulations every day presents unique challenges and opportunities. At this scale, it is nearly impossible to manually go through thousands of test results to look for metrics that lead to actionable insights. There are also huge computing and storage costs associated with running such a large amount of tests. Using data science, we can more efficiently analyze and understand test results and optimize system performance.
How did you go from Google to self-driving technology?
Before joining Aurora, I worked on Google Search for 18 years, most recently leading the Data Science & Engineering team for Google Search & Assistant. When I first started thinking about joining Aurora, it reminded me of a feeling I got when I first joined Google—an excitement about the possibility of making a big impact in a transformative industry. I spoke with Chris Urmson and was blown away by his vision to deliver the benefits of self-driving technology safely, quickly, and broadly, and the rest is history.
What is your leadership philosophy?
I play golf with my husband and kids, and I like to say that, like golf, work and parenting take a lot of patience and consistency. It’s not really a game against others, but a game against yourself. I try to always celebrate the small wins and remember my good shots, but I don’t keep score. Golf makes me a better leader and parent.
Who inspires you?
My mom has always been my biggest inspiration. She is the most resilient and loving person I’ve ever met despite having a challenging upbringing. When she was eight years old, overnight, she lost her family and all their possessions and grew up in an orphanage. Later during China’s Cultural Revolution, she was sent to the countryside to get re-educated. Despite these hardships, she secured a college education, cultivated her own successful career as a veterinarian, built a successful dairy processing plant, and raised two daughters who together have two PhDs and a JD.
Through these experiences, my parents impressed upon me and my sister the importance of an education and having a career. Their stories inspired me to do hard things and carve my own path in the technology industry, and I’m passing on those lessons to my own children.