If Ben Zabihi had started his career five years ago, his workday would have looked very different than it does today.
A few years ago, he might have spent much of his time formatting code and writing documentation. Now Zabihi, who has been working as a software engineer at a small New York City startup since December, said a good portion of his day is spent using AI tools — not just to write code, but also as a research assistant to better understand his industry and business terminology.
The 23-year-old entered the profession at a time when companies and workers are actively testing and debating the extent to which AI is helpful and what still requires a human touch.
Though Zabihi said that relying too much on AI at the start of his career could result in a weaker foundation for his learning in the long term, he also knows he has to use the technology and optimize his workflow.
Read more about software engineers and AI
The tasks that used to keep entry-level engineers busy might not be as important as they once were, he said.
Instead, he’s focusing on the bigger picture: work like understanding business goals, system architecture, scaling, and security risks, which once were the domain of more senior engineers.
Risk and opportunity
While many recent grads see AI as a way to gain superpowers quickly, some industry veterans worry that the technology erodes a formative stage of learning that builds judgment and problem-solving skills — a gap that may only become clear as today’s engineers advance.
When 36-year-old engineer Georgian Tutuianu entered the field a few years ago, he said 95% of the job was painful. For today’s junior engineers, though, there are many shortcuts — and he’s worried those can come at the expense of deeper understanding.
For example, a core part of his job is managing codebases through pull requests, where engineers submit code for review before it’s merged into the system. Tutuianu said he used to review around 100 to 500 lines of code in a pull request. Now, with LLMs, it’s easily over a thousand, and he said he sees workers often add layers of complexity they don’t understand.
“It’s super concerning because then you have just a pile of terribleness that you have to contend with,” Tutuianu said. “It’s literally just pollution.”
He said he worries that junior engineers may be outsourcing the hardest part of the job — wrestling with what they don’t understand — to LLMs.
Zabihi and Tutuianu’s differing experiences reflect a wider shift in the industry. As one of the fastest sectors to adopt AI, software engineering is being transformed — and entry-level roles, which were once the training grounds for mastering the complexities of the job, are fundamentally changing.
With that comes risk, but also opportunity.
Getting ahead
There’s no crystal ball to predict where the industry is headed, but one thing is clear: Junior developers are navigating a murky employment market as the industry undergoes a tectonic shift. That means they’ll need to move quickly to stay relevant.
The shift in focus may also force a rethinking of the fundamentals of the job. If AI can handle much of the code itself, the value of an engineer might lie less in perfecting syntax and more in gaining a broader expertise in defining problems and architecting solutions.
“The question then is, how do the requirements of the job and the skills change?” Matt Kropp, managing director and senior partner and chief AI officer of BCG X, the tech division of Boston Consulting Group, told Business Insider. “If you’re a junior engineer, how do you make sure that you meet those skills in the market?
Keith Ballinger, Google’s vice president and general manager of Developer & Experiences, told Business Insider that “nothing beats doing it.”
“You don’t need to ask for anybody’s permission to do something significant and meaningful,” Ballinger said. “Just put together a cool app and post it on a website.”
Ballinger said that most software engineers didn’t enter the field to write code in a specific language or framework. A developer’s job is to use technology to solve problems and apply engineering techniques, he said. Great engineers have always known how to break down problems into smaller ones, and now agents can help handle the rest, Ballinger said.
“That’s a skill that we can teach and that people can pick up, but now it’s more important than ever, and certainly more important than memorizing how an API works,” Ballinger said.
As entry-level hiring opportunities shift, Mohit Bhende, the CEO and cofounder of engineering hiring platform Karat, said aspiring engineers should seek out organizations committed to training junior talent. Those opportunities may increasingly lie outside traditional tech, he said.
Bhende said he expects more talent to move to sectors like finance and healthcare, where AI adoption is slower, and security concerns elevate the value of human oversight.
He said CTOs are also increasingly seeking engineers who understand the business side of their work. Bhende said that aspiring engineers should prioritize developing domain knowledge, whether through on-the-job training or formal education.
“Maybe you’re graduating not just with the computer science degree, but you’re graduating with that, plus a business degree,” Bhende said, adding that he thinks “the jobs of the future are going to merge those two.”
Zabihi, for one, is bullish about what the rapidly evolving tech will mean for his career. He said his output is significantly higher because of AI — and ultimately, that’s what he’s being paid for.
“As a junior dev, you’ve never gotten a better bang for your buck,” Zabihi said.
