Watch Drive AI's Deep-Learning Car Zip Through the Rain, the Mountain View-based autonomous car startup, has been working hard. The company is using deep-learning to teach its vehicles how to drive — meaning that it can learn how to deal with situations, develop its own rules, and apply those rules to unfamiliar situations. It’s advanced stuff, and on Tuesday, gave the world its first glimpse at its technology in action.

Driving down a rain-soaked nighttime Mountain View street, the team put the machine through its paces in a four-minute video that shows the car managing reflections on the wet street could have confused the system, as could a broken red light, but dealt with all these obstacles effortlessly.

The startup, which officially launched in August 2016, is operating in a crowded space, with big names like Tesla and Google spin-off Waymo also looking for a slice of self-driving pie.

However, has a lot of resources to its name. Before launching, it closed a $12 million funding round. The startup’s also one of 22 entities that have an Autonomous Vehicle Testing Permit from the state of California and, judging by the video, it’s aiming to get as much use out of it as possible.

“’s unique deep learning technology processes information and makes decisions more like the human brain does,” the company announced in a Medium post. “Our technology allows us to scale to all kinds of challenging environments.”

When the team launches a product, it’s expected to take the form of a kit that can work with existing cars. The kit is also expected to feature a communication system, capable of displaying emojis and other messages to passers-by. The rationale is that communicating with pedestrians will help others understand how the car intends to act next, stopping them from putting themselves in harm’s way.

Watch the technology in action here:

And if you think you have the skills to work for them,’s “Careers” page has eleven openings, from intern up to software engineer.

Related Tags