Learning new AI skills has become like a second job for software engineers

• Many developers are dedicating hours a week to experimenting with AI tools and taking new courses.
• Software engineers told Business Insider that upskilling is required to stay relevant in the AI era.
• Some said they spend around 20 hours weekly learning new AI-related topics.

Software engineers are in the midst of a sweeping industry shift — and they've picked up a second job along the way: upskilling.

Some developers are deliberately pivoting into AI-focused roles at the cutting edge of the technology. However, even those who remain in traditional engineering roles say keeping up with the pace of change can feel daunting.

"In order to keep up with everything that's coming out right now, you have to have no job," Jason Grad, the cofounder and CEO of tech startup Massive.

He told Business Insider that he spends roughly 20% of his time learning, reading, and testing out new things.

In conversations with Business Insider, four other developers and three tech leaders described carving out anywhere from an extra hour each day to as much as 20+ hours a week to learn new tools and adapt to rapidly changing demands.

While the amount of time dedicated to upskilling varied among engineers, the pressure they all described feeling was the same: Get on board or fall behind.

A new requirement

In addition to constantly learning on the job, Pratiksha Patnaik, a 30-year-old infrastructure engineer at Google, said AI has changed the scope of her work, shifting from networking security and infrastructure to helping customers adopt generative AI products and solutions. She spends a couple of hours a week learning new AI concepts and said she'd "be left behind" if she didn't.

"It is required as part of my role," Patnaik told Business Insider.

Maahir Sharma, a 24-year-old software engineer, has spent over a decade refining his coding skills, and said he initially felt concerned about AI taking his job, as the models rapidly improved a few months ago.

However, he said he now feels optimistic about his future as a developer because he's spending time learning about new concepts and educating himself on AI.

"Backend engineering is now going to transition into AI engineering," Sharma, who said he spends about 20 hours a week upskilling, told Business Insider.

In addition to experimenting with tools, Sharma said he's been taking online courses about topics like prompt engineering, building out multi-orchestration AI agents, and how model context protocol servers work.

Marc Kriguer, a lead software engineer at an insurance company, told Business Insider that he spends between 30 minutes and an hour each day after work keeping up with new tools.

He said he uses CodeSignal, an AI-powered platform that offers one free token every 24 hours. That token allows him to practice Go, a programming language, he said. He has also spent time taking AI courses and improving his prompting skills.

Beyond hands-on experimentation and courses, part of the challenge is figuring out what's next. Feneel Doshi, a 27-year-old software engineer at a Jersey City startup, said he spends four to five hours a week experimenting with tools. He starts his weekday mornings by reading newsletters like TLDR and Flipboard, and developer blogs from OpenAI and Anthropic, to keep tabs on what's new in the industry.

Individual contributors aren't alone in the scramble. Tech leaders who are expected to lead the charge on AI adoption have also said they're carving out extra hours each day to keep pace with the change.

Google VP and general manager Keith Ballinger told Business Insider that he sends out weekly email tracking what he's been working on to keep himself accountable. He said he spends over 20 hours a week experimenting with tools.

The joy of experimentation

Learning new skills doesn't always feel like a grind, several engineers said.

"Once you actually see something going live in production and you see the results of it, it kind of becomes addictive at that point," Sharma told Business Insider.

Sharma added that building projects for monetary gain doesn't feel as rewarding, and part of the fun is that AI experiments done in your own time.

Ballinger similarly said he loves the new era of AI engineering and feels there's never been a better time to be a developer. He said he doesn't spend 20+ hours a week experimenting with AI because he has to. He does it because he enjoys it so much.

There's always the possibility you could be creating the next big tool or feature.

Emrick Donadei, a 32-year-old software engineer at Google who transitioned into an AI role, said part of his excitement in participating in a hackathon was the possibility that what he was building could become a real feature.

Sharma pointed out that tools like OpenClaw were built in one month and went viral. That's something that he feels "optimistic" about when he builds, he said.