Ethan Sorrelgreen '99

Chief Product Officer, CARMERA

Driving a car safely across town requires a high level of brain activity. The onslaught of sensory information—colors into optical neurons and sounds into auditory neurons—must be absorbed and contextualized, then sent to another part of the brain that delivers a command to turn right or left or stop. If navigating in a familiar neighborhood, your brain is rather nimbly performing something called “localization,” which means you do not need to see any signs or look at a map to know the street network and understand where you are.

A self-driving car must replicate every one of these neuronal functions. This daunting task is why Ethan Sorrelgreen ’99 gets up every morning. He is the chief product officer for CARMERA, a start-up company based in Seattle and New York City that is building a real-time, street-level intelligence platform to create maps that power autonomous vehicles.

“I’m the bridge between the customers and what they need and the engineers and what they build,” says Ethan, who founded CARMERA’s Seattle office a few years ago. Before working at CARMERA, he was an engineering manager at Apple and Amazon, and built both of their v1 Maps Applications.

Operating in a World of Obstacles

At CARMERA, Ethan is responsible for “mapping,” an essential element for self-driving cars to navigate safely. In a world of obstacles, the car needs maps that constantly refresh and deliver important information— such as a sudden detour due to construction, or a double-parked truck—so it can take the safest and most efficient route possible.

“Driving conditions for a self-driving car have to be in real time, not a month ago,” says Ethan. “You have to keep that info really up to date at a rate of high efficiency.”

From Pulsars to Map Data

Historically, Ethan has always been passionate about mapping. When he played on the basketball team at The Northwest School, he was always at the front of the bus telling the coach how to navigate through the streets to the games.

After graduation, Ethan headed to Haverford College where he majored in astronomy and minored in physics. There, he studied pulsars (old dead stars that have collapsed and become neutron stars)—which spin at a fantastic rate. One pulsar he studied had a spin rate of 16 milliseconds.

“It takes a huge amount of compute time to find the signal of that 16 millisecond pulsar in dataset spanning 12 hours,” explains Ethan. It would take at least 10 computers processing all semester long in order to complete the project but only one of them had a hard drive large enough to store the data. His solution was to build a “Linux cluster” to break the job up over a number of machines in order to study that pulsar. “This was the beginning of my experience working with distributed systems,” reveals Ethan. “We use similar processes today to build map data in real time.”

Ethan followed a circuitous route to the field of big tech and mapping. After college, he first pursued a career in sales and recruiting. He also worked at a research institute, and then at a startup in the public bus prediction space before winding up at Apple.

This profile originally appeared in the Fall 2018 issue of The Northwest School Magazine.