Many roads lead to a career in the self-driving car industry. That painfully obvious pun is actually one of the truest things you can say about this nascent, multidisciplinary enterprise, and it also encapsulates the challenge educators who want to prepare their students to work in this industry are facing today.
“There’s a misconception that there’s such a thing as ‘self-driving car technology,’” says Sudha Jamthe, who designed and teaches the first autonomous vehicles business course for Stanford University’s Continuing Studies program. “But the self-driving car has many, many layers of tech—of software, of engineering, and design—which is the beauty of it. People working in this industry have a wide range of skills and experience.”
Predictions about when self-driving cars will take over our streets and highways vary widely, but Alexander Wyglinski, professor of electrical and computer engineering at Worcester Polytechnic Institute (WPI) in Massachusetts, says these yet-to-go-mainstream machines are generating real opportunities today for job seekers with the right combinations of qualifications.
“What people in this industry are looking for depends a lot on the company,” Wyglinski says. “There are the Teslas, Waymos and Ubers of the world—companies looking to get self-driving vehicles on the road in two or three years. There are the startup companies trying to figure out what the next generation is going to look like. And there are companies working on systems that will go into self-driving cars—new radar and lidar systems, new decision-making algorithms and software, and new mechanical components for these vehicles.”
One thing all of those companies seem to have in common, says John M. Dolan, principal systems scientist at Carnegie Mellon University and a self-driving vehicle researcher in the school’s Robotics Institute, is the value they place on employees with that broad skill set.
“Generally speaking, it’s not enough to be just a programmer or an engineer,” Dolan says. “You might be hired to write the code or design the hardware, but they expect you to be able to understand the big picture.”
All three experts interviewed for this story emphasized that the self-driving car/autonomous vehicle industry is still in its infancy, and that it is evolving quickly and often in unpredictable ways. With that caveat in place, they offered the following advice for students and educators:
Take A Lot of Math
Students would be well served to take every math class they can get in high school, and then continue with advanced courses in college. Especially important: linear algebra and calculus. “Linear algebra is at the heart of how the car learns to drive itself,” says Jamthe.
Learn to Program More Than Just Python
Python is the most commonly taught language in secondary schools, and it’s the language used in TensorFlow, the popular open-source machine learning library—which is good. But even better, says Dolan: C++. “It’s the go-to language for fast execution and building substantial architectures,” he says. “Python doesn’t do as well when it comes to dealing with real-time systems, where you’re trying to make sure you have millisecond timing on different events and things like that. That’s where C and C++ bring in the structural part, which object-oriented programming is very good at.”
Join the Robotics Club
All three offered this recommendation enthusiastically. A self-driving car is essentially a robot, so getting experience working with these machine will be invaluable in the job market, they said. Robotics teams include builders and coders—and promoters—so they offer a kind of industry preview. “The more exposure to whatever elements are available in this area, the younger the better,” Dolan said.
Study Machine Learning
Most colleges offer classes in artificial intelligence, machine learning, and deep learning, but high schoolers will likely have to look for classes outside school, Jamthe says. “There’s all this fluffy talk about AI and it’s so scary, but what people are really learning is machine learning,” she says. “I think it’s something young students need to learn, and they can learn it.”
Study Probability and Statistics
This is another key subject, says Dolan. “Take it in high school if you can,” he says. “You’ll find it gets applied to localizing robots and high-def mapping. They use SLAM techniques (simultaneous localization and mapping), which depends on probability and statistics.”
Take Some Art Classes
Once we establish vehicles that truly drive themselves, their interiors are going to undergo a serious redesign. Having some familiarity with design will help you think outside the box, Jamthe said. “It’s going to be a highly competitive field, and you’ll need creativity to differentiate yourself, both as an employee and potentially an entrepreneur.”
Pursue Multiple Majors
The self-driving car industry is multidisciplinary, and the students who want to work in it should be too, says Wyglinski. “We see a lot of dual majors here at WPI—combinations like mechanical and electrical engineering, or computer science and electrical engineering,” he says. “And our robotics engineering program is inherently multidisciplinary. So, we’re graduating students already with the ability to look at a problem from multiple perspectives, and that’s a key skill in the self-driving car industry.”
Work on Your Own Car
“I say this to people, and I get a few looks, but if you think about it, it’s not such an odd recommendation,” Wyglinski says. “If you want to work in an industry focused on cars, you should take a look at one. Change your oil. Change your brake pads. Go in with the mechanic when he or she has your car up on the hoist and look around. It’s about seeing where theory will meet practice. And it’ll make you less afraid to jump in and play around with physical systems.”
Never Stop Learning
“Keep in mind that this industry is evolving very quickly,” says Jamthe. “The jobs that are out there today are going to be different tomorrow, and it’s very hard to predict what they will be. So, general knowledge, as opposed to specialized knowledge, is likely to be very important in the long run.”
