When cybersecurity firm Malwarebytes started automating its quality assurance testing last year, it knew that the move could put more than four dozen employees out of work.

Rather than simply replace them with tech, though, the Santa Clara, Calif., firm turned to tech to save their jobs.

The company signed up with Udemy, an online learning platform that teaches courses from data science to sourdough bread-making. Malwarebytes identified the skills its quality assurance testers would need to stay relevant in the rapidly changing cybersecurity industry.

Then it told its staffers to buckle in — it was time to get “up-skilled.”

As automation becomes ubiquitous, education startups such as Udemy, Coursera and General Assembly are positioning themselves as the nexus between today’s workforce and tomorrow’s jobs.

Unlike traditional college programs that can take anywhere from two to four years and tens of thousands of dollars to complete, the online schools frame themselves like vocational programs for the Silicon Valley set. They offer training in a specific skill — say, learning a programming language — in months, rather than years. Coursera offers some of its courses for free but also has programs where it charges a few hundred dollars. Udemy’s courses for individuals start at $10.99, while its enterprise-facing arm, Udemy for Business, offers group rates for companies that want to sign up entire teams.

Online learning and micro-certification are increasingly appealing to individuals and businesses seeking to adjust to automation, or introduce automation without having to fire and rehire a more highly skilled workforce.

“It’s completely changed how we think of training,” said Brent Boeckman, global learning and development manager at Malwarebytes. “In the past, you’d have to budget to send employees across several states to do a course, you’d have to think about travel and hotel costs.”

With automation predicted to displace as many as 73 million, or a third, of U.S. workers by 2030, futurists and automation experts say that retraining at every stage of a person’s career will probably be the most viable path for those who intend to keep working.

The quality assurance testers at Malwarebytes, for example, previously tested the company’s software manually: They would run a program, check to see if it logged any errors, lather, rinse, repeat. A computer now handles those mundane tasks, which frees the “up-skilled” workers to take on the more complex tasks. With their newfound programming knowledge, they’re able to identify errors in the company’s code base and contribute to a solution with the company’s developers.

In San Francisco, bonds trader David Gagnon, 50, has done more than a dozen courses in programming and machine learning on Coursera to protect his own career. In his more than 20 years in the business, he’s seen global trading desks go from employing 10 traders and 10 assistants, to 10 traders and a handful of computer programmers who, working together, do 10 times the trades.

Automation is far from perfect, Gagnon said, which is why having basic programming and data science skills lets him spot errors and vulnerabilities before they balloon and cause mayhem — making him a more valuable bonds trader.