AI has upended what it means to be a software program engineer. It is also altering the way you turn out to be one.
Earlier than the AI growth, software program engineering interviews revolved round writing code from reminiscence, fixing algorithmic puzzles, and demonstrating technical velocity underneath stress.
At this time, CEOs are scouring GitHub and X for undiscovered expertise, encouraging candidates to make use of AI throughout interview rounds, and more and more hiring for judgment and style.
For the tens of millions of software program engineers all over the world, it means navigating a radically completely different job market from only a few years in the past.
Le’ale Addison, a current laptop science graduate who interned at Amazon and KPMG earlier than touchdown a know-how internship at packaging firm Smurfit Westrock, has skilled this transformation firsthand.
Karen Haberberg for BI
Two years in the past, the 22-year-old stated she was required to share her display throughout technical assessments to make sure she wasn’t dishonest. By her senior 12 months, Addison stated she might brazenly use AI or Google to seek for responses in actual time.
Throughout her most up-to-date job search just a few months in the past, employers peppered the interviews with questions on AI: How acquainted was she with machine studying? What about pure language processing? How was she utilizing AI in her present workflow — and the way would she apply it to the brand new job?
“These questions weren’t beforehand requested,” Addison stated.
Addison’s expertise displays a broader recalibration enjoying out throughout the tech world, with corporations like Dropbox and Cisco asking engineers to show their AI expertise in the course of the hiring course of.
These seeking to break into an business in flux or discover their subsequent alternative face a aggressive market. Whereas hiring is up, 74% of builders are struggling to land jobs, based on a 2025 report from HackerRank.
Enterprise Insider spoke to tech corporations, AI startups, and profession coaches to reply one overarching query: What does it take to get a coding job when AI is writing many of the code?
Goodbye résumés, howdy GitHub
Large Tech corporations, together with Google and Meta, have been battling it out for high AI engineering expertise, providing eye-watering compensation and entry to compute.
In the meantime, startups have gotten inventive to woo potential hires. Executives at AI coding startups Cognition, Base44, and Replit say they frequently scout engineers by trawling X posts and GitHub pages, two platforms the place engineers submit their newest initiatives to be lauded — or picked aside — by friends.
At Replit, X is shortly changing into a “fundamental medium” for recruiting, stated the corporate’s chief folks officer, Stacey La Torre. They’re getting rank-and-file staff in on the hunt, too.
Karen Haberberg for BI
“We now have a Slack channel referred to as ‘Expertise Spot’ for folks to mainly say, ‘Hey, I linked with this individual on LinkedIn, or I met them at a convention or on X,” La Torre stated.
Emily Cohen, Cognition’s head of individuals and operations, stated that each worker on the firm sees discovering expertise as a part of their duty and that she has flown all around the world herself to persuade good engineers to hitch the startup.
“Simply final week I drove a candidate to the airport as a result of I needed to be the final individual they talked to earlier than they left San Francisco,” Cohen advised Enterprise Insider.
AI takes over the technical interview
It isn’t simply the place corporations are discovering expertise. Vibe coding has modified how employers conduct interviews — and what they need to hear from candidates.
LeetCode, a web-based platform used to guage candidates’ expertise, was as soon as a ceremony of passage for candidates. Now, it is simply the primary hurdle, if it is included in any respect.
Xavier Contreras, a New York Metropolis-based head of information engineering at a hedge fund, has been constructing, designing, and sustaining programs for a couple of decade. His most up-to-date job search, he stated, seemed vastly completely different from those he went by way of 5 or 6 years in the past.
Again then, interviews revolved round coding challenges, he stated. In his most up-to-date job search, he was requested to stroll interviewers by way of his reasoning, similar to explaining the structure behind initiatives he is constructed or defending technical choices. When he was given take-home assignments, he was allowed to make use of AI for work that may have beforehand taken a month. With AI, the anticipated timeline was three days.
Karen Haberberg for BI
Now that AI can do extra of the grunt work, employers need to know that their employees perceive what the tech is doing — and might appropriate it when it goes off-course.
Probably the most precious technical expertise immediately are programs pondering, drawback fixing, and the power to function new agentic programs shortly, stated LinkedIn’s head of world expertise acquisition, Erin Scruggs.
“Corporations want individuals who can translate complexity, make choices about how programs ought to behave, and lead groups by way of speedy technological change,” Scruggs stated.
That is main some corporations to assign assessments that provide a broader view of a candidate’s ability set. Scott McGuckin, VP of world expertise acquisition at Cisco, stated the corporate is shifting from technical coding challenges to project-based workouts. The corporate can also be integrating AI-assisted improvement environments into interviews to watch how candidates function in an AI-enabled workflow.
“The human ingredient of oversight and experience is extra essential than ever,” he stated.
Trying to find ‘information unicorns’
The shift towards AI-assisted hiring assessments does not imply technical rigor is not vital. Sundeep Teki, a profession coach who helps place expertise at frontier AI labs, stated that for roles at coveted corporations like OpenAI and Anthropic, the technical bar is continually evolving and getting larger.
Along with the technical know-how, AI savviness, and gentle expertise required to excel on the job, corporations are putting better worth on tradition match. He stated that Anthropic, for instance, features a devoted non-technical tradition interview for each function. He stated that if a candidate does not appear to gel with the corporate’s mission, they usually will not make the reduce.
One other method to inform if somebody is a match in your workforce’s tradition? Ask them to start out doing the work.
Work trials, the place candidates present up in individual and work with their potential workforce, are gaining popularity. AI coding startups like Lovable, Cursor, and Kilo all conduct trials as a means for each candidate and firm to see if it is a good match.
“You’ll meet us all in Amsterdam and begin in individual, and also you’re anticipated to be transport product that options the subsequent day,” CEO Scott Breitenother stated about Kilo’s newest abroad bootcamp. “Numerous of us will take PTO off their present job, come to work with Kilo for our focus week, after which in the event that they make the reduce, they’re going to give up their job.”
A part of the explanation the hiring course of seems to be completely different is as a result of the roles themselves have modified. Within the age of AI, the place many features of a developer’s conventional work have been automated, technical roles are mixing, and employers are in search of candidates with a broader, extra versatile ability set — a jack-of-all-trades who can adapt throughout features.
Contreras, the hedge fund information engineer, stated that software program engineering, information analytics, and information science have all moved underneath the identical umbrella. Which means completely different skillsets are wanted to get the job — and to retain it.
“They need to rent one individual to do the function of three,” Contreras stated.
Whereas employers appear to be frequently refining what they’re in search of, corporations are looking for what Contreras calls “information unicorns.”
“It is much more hectic,” he stated. “There’s simply a lot extra data to know now.”