AI has upended what it means to be a software engineer. It's also changing how you become one.

Before the AI boom, software engineering interviews revolved around writing code from memory, solving algorithmic puzzles, and demonstrating technical speed under pressure.

Today, CEOs are scouring GitHub and X for undiscovered talent, encouraging candidates to use AI during interview rounds, and increasingly hiring for judgment and taste.

For the millions of software engineers around the world, it means navigating a radically different job market from just a few years ago.

Le'ale Addison, a recent computer science graduate who interned at Amazon and KPMG before landing a technology internship at packaging company Smurfit Westrock, has experienced this transformation firsthand.

AI has become so proficient at writing code that it's now commonplace in technical interviews. Karen Haberberg for BI

Two years ago, the 22-year-old said she was required to share her screen during technical assessments to ensure she wasn't cheating. By her senior year, Addison said she could openly use AI or Google to search for responses in real time.

During her most recent job search a few months ago, employers peppered the interviews with questions about AI: How familiar was she with machine learning? What about natural language processing? How was she using AI in her current workflow — and how would she use it on the new job?

"Those questions weren't previously asked," Addison said.

Addison's experience reflects a broader recalibration playing out across the tech world, with companies like Dropbox and Cisco asking engineers to prove their AI skills during the hiring process.

Those looking to break into an industry in flux or find their next opportunity face a competitive market. While hiring is up, 74% of developers are struggling to land jobs, according to a 2025 report from HackerRank.

Business Insider spoke to tech companies, AI startups, and career coaches to answer one overarching question: What does it take to get a coding job when AI is writing most of the code?

Goodbye résumés, hello GitHub

Big Tech firms, including Google and Meta, have been battling it out for top AI engineering talent, offering eye-watering compensation and access to compute.

Meanwhile, startups have gotten creative to woo potential hires. Executives at AI coding startups Cognition, Base44, and Replit say they regularly scout engineers by trawling X posts and GitHub pages, two platforms where engineers post their latest projects to be lauded — or picked apart — by peers.

At Replit, X is quickly becoming a "main medium" for recruiting, said the company's chief people officer, Stacey La Torre. They're getting rank-and-file employees in on the hunt, too.

Xavier Contreras, head of data engineering at a hedge fund, has been building, designing, and maintaining systems for about a decade. Karen Haberberg for BI

"We have a Slack channel called 'Talent Spot' for people to basically say, 'Hey, I connected with this person on LinkedIn, or I met them at a conference or on X," La Torre said.

Emily Cohen, Cognition's head of people and operations, said that every employee at the company sees finding talent as part of their responsibility and that she has flown all over the world herself to convince good engineers to join the startup.

"Just last week I drove a candidate to the airport because I wanted to be the last person they talked to before they left San Francisco," Cohen told Business Insider.

AI takes over the technical interview

It's not just where companies are finding talent. Vibe coding has changed how employers conduct interviews — and what they want to hear from candidates.

LeetCode, an online platform used to evaluate candidates' skills, was once a rite of passage for candidates. Now, it's just the first hurdle, if it's included at all.

Xavier Contreras, a New York City-based head of data engineering at a hedge fund, has been building, designing, and maintaining systems for about a decade. His most recent job search, he said, looked vastly different from the ones he went through five or six years ago.

Back then, interviews revolved around coding challenges, he said. In his most recent job search, he was asked to walk interviewers through his reasoning, such as explaining the architecture behind projects he's built or defending technical decisions. When he was given take-home assignments, he was allowed to use AI for work that would have previously taken a month. With AI, the expected timeline was three days.

Contreras said that AI has moved software engineering, data analytics, and data science under the same umbrella. Karen Haberberg for BI

Now that AI can do more of the grunt work, employers want to know that their workers understand what the tech is doing — and can correct it when it goes off-course.

The most valuable technical skills today are systems thinking, problem solving, and the ability to operate new agentic systems quickly, said LinkedIn's head of global talent acquisition, Erin Scruggs.

"Companies need people who can translate complexity, make decisions about how systems should behave, and lead teams through rapid technological change," Scruggs said.

That's leading some companies to assign tests that offer a broader view of a candidate's skill set. Scott McGuckin, VP of global talent acquisition at Cisco, said the company is moving from technical coding challenges to project-based exercises. The company is also integrating AI-assisted development environments into interviews to observe how candidates operate in an AI-enabled workflow.

"The human element of oversight and expertise is more crucial than ever," he said.

Searching for 'data unicorns'

The shift toward AI-assisted hiring tests doesn't mean technical rigor is no longer important. Sundeep Teki, a career coach who helps place talent at frontier AI labs, said that for roles at coveted companies like OpenAI and Anthropic, the technical bar is constantly evolving and getting higher.

In addition to the technical know-how, AI savviness, and soft skills required to excel at the job, companies are placing greater value on culture fit. He said that Anthropic, for example, includes a dedicated non-technical culture interview for every role. He said that if a candidate doesn't seem to gel with the company's mission, they often won't make the cut.

Another way to tell if someone is a fit for your team's culture? Ask them to start doing the work.

Work trials, where candidates show up in person and work with their prospective team, are becoming more popular. AI coding startups like Lovable, Cursor, and Kilo all conduct trials as a way for both candidate and company to see if it's a good fit.

"You will meet us all in Amsterdam and start in person, and you're expected to be shipping product that features the next day," CEO Scott Breitenother said about Kilo's latest overseas bootcamp. "A lot of folks will take PTO off their current job, come to work with Kilo for our focus week, and then if they make the cut, they'll quit their job."

Part of the reason the hiring process looks different is because the jobs themselves have changed. In the age of AI, where many aspects of a developer's traditional work have been automated, technical roles are blending, and employers are looking for candidates with a broader, more versatile skill set — a jack-of-all-trades who can adapt across functions.

Contreras, the hedge fund data engineer, said that software engineering, data analytics, and data science have all moved under the same umbrella. That means different skillsets are needed to get the job — and to retain it.

"They want to hire one person to do the role of three," Contreras said.

While employers seem to be continually refining what they're looking for, companies are searching for what Contreras calls "data unicorns."

"It's a lot more hectic," he said. "There's just so much more information to know now."

What do you think about how the software engineering industry is changing? Let us know below: