Carolina Public Press–The depleted job market for new graduates of computer science has garnered national attention, with Federal Reserve Bank of New York data showing that computer science graduates are among the highest rates of unemployment at 6.1%.
The New York Times reported that with major companies like Microsoft and Amazon laying off workers and instead opting to implement artificial intelligence for its coding properties could be the culprit of the bleak market.
North Carolina’s top three universities in the Research Triangle all sit in the U.S. News and World Reports’ list of best schools for computer science with Duke University being ranked 25th, UNC-Chapel Hill ranked 27th and NC State University ranked 51st.
How are these universities responding to the challenges their students face as they head into the work force?
A struggling market
Casey Dockery graduated from NC State in 2024 after just three years with a bachelor’s degree in computer science and a minor in graphic communications.
He’s been looking for a job in his field for close to 16 months now, and he typically filters for jobs pertaining to web and software development. While he said a remote job would be nice, he’s open to any opportunities in the field.
When Dockery was nearing graduation last year, he had already prepared himself for the reality of the market. His fears were unfortunately validated.
“Before I started my last semester, I pretty much knew that the job market was going to be really bad,” he said.
“So I put in a lot of applications for jobs that I would have liked to have gotten like the web developer positions, but I knew those would be competitive. And it was true, as I never got a single interview for a very long time.”
He decided to look for work at companies that didn’t necessarily have computer science positions available at the time but could have some open up later. This strategy would get his foot in the door somewhere adjacent to his desired career in the meantime, he said.
He started working as a material handler at a biotech company, where he has stayed for eight months. Shortly after hiring him, the company gutted the IT department and stopped hiring for computer science jobs altogether.
Getting an interview is the hardest part, Dockery said. Besides the one needed for his current job, he’s gotten one interview for a computer science job since graduating.
He had two referrals for that company, so he felt his chances were better than most. He made it through three rounds of interviews, which took three weeks to complete. In the final round, Dockery felt blindsided by the difficulty of the questions for a job that was supposedly entry-level.
“It seemed to me that the requirements for the job position were very light,” Dockery said.
“But then whenever I actually got to the final round, they started asking difficult questions — higher than entry-level. And then I even asked the people that work there currently in the position I applied for, and they said that they didn’t even know the answers to those questions.”
Dockery decided to study computer science at the age of seven, when he learned his interest in computers was something that could be an actual career, not necessarily because it offered job security or a comfortable salary.
When he started at NC State in 2021, he was constantly validated by his peers who said it would be easy for him to find a job, that he would be set for life. He felt confident in his career path but still worked hard to get an internship while in school, though he wasn’t able to land one.
As he began his third year in 2023, the year he was set to graduate, Dockery felt a tone shift.
“I realized that it wasn’t going to be easy, because I saw how the computer science job market was,” he said.
“Videos on Instagram, like making fun of computer science majors going to be homeless, stuff like that. So yeah, starting out, I definitely thought that it was going to be a lot easier than it was.”
AI’s impact on computer science
Like many, Dockery believes AI is playing a significant role in the direction the computer science market is heading — specifically the entry-level roles.
“It seems like mid-level and seniors especially aren’t as impacted. It seems like it’s not that hard to transition from job to job once you get a little bit of experience,” he said.
“Just getting your foot in the door is the most difficult part. And obviously, AI would mostly impact entry-level jobs because it’s the easiest work and it’s the most easily replicated by AI.”
Generative AI tools like ChatGPT are able to write and debug code, some of the most basic requirements of computer scientists.
Microsoft’s Copilot, touted by the tech giant as “Your AI Companion,” is specifically designed for writing code and answering coding-related questions. Claude is even more involved, coding even what Dockery called the creative aspects of code, not just the technical.
Sarah Heckman, Director of Undergraduate Programs at NC State’s Department of Computer Science, said she thinks of AI as a way to enhance the developer’s experience and increase productivity. Ultimately, the engineer still needs to understand coding and debugging techniques on their own.
The most important thing about coding is the creation of value. That’s where the human on the other end of the computer is necessary, because it’s not the code itself that provides value but the way the code is put together.
“There’s still that human in the loop because the AI is not going to create it on its own,” she said.
“You’re going to have to prompt the AI tool to provide some of that feedback for you and then integrate it into a working system. So enhance, not replace, is the way I like to think of it.”
AI can also support software engineering skills through helping the engineer understand the design and explore its requirements.
Jian Pei, chair of the computer science department at Duke University, said coding still requires people in order to guide the process and, essentially, make the computer work. But instead of coding line-by-line, AI speeds the process.
That’s the most important change overall, and students need to be able to use those tools to their advantage in order to be the most efficient, he said.
“We should not abandon the students and say, ‘Oh, hey students, you cannot use AI to do your homework,’” he said. “Instead we should encourage the student to think about what is the best way to use AI tools to help you to do the homework.”
Universities adjust
Computer scientists see the rapid development of AI as both a challenge and an opportunity, Pei said.
While it is true that computer scientists whose sole skill is the traditional form of coding may struggle to find work as AI becomes more prevalent, Pei said AI has other places within the discipline that can create more jobs such as machine learning and data science.
To adjust to this changing landscape and in the interest of preparing students for AI’s various opportunities, Duke has upgraded its curriculum to include new and more material related to AI.
In 2026, the department will launch a new class called Mathematics for Artificial Intelligence, which Pei said will prepare students to tackle AI more quickly and effectively, rather than taking several courses before ever interacting with the real thing. Duke has also offered a concentration in AI and machine learning for the Bachelor of Science track for several years.
“We foresee this will be an important opportunity for those students, but they may not have time to go through such a long path in the traditional way, to learn the math, learn the coding,” he said.
“And we want to sharpen them so that they can touch the frontier even earlier. We already see the outcome. For example, many of our students at the very beginning of this year worried whether they could find an internship, but it turned out, many of them found a wonderful internship. They are coming back to campus now, and many of them told me they got return offers.”
Heckman said AI has changed how the department is approaching education in some ways. The university wants to educate computer scientists and software engineers who are prepared to address challenges and ensure they have the tools to be successful, and sometimes that includes AI.
Over the summer the department has studied various AI tools and considered the ways it wants to integrate them into coursework. Professors consider the learning outcomes of their course and what they want their students to be able to accomplish at the end of the semester to determine which tools are appropriate
“In some classes, we might be more restrictive on the use of AI because we want to make sure our students know how to do fundamental programming skills so that they can then understand what these AI tools are producing when they become professional software engineers,” she said.
“But in later coursework, we know that students are going to be using these AI tools professionally, and we want them to start using them professionally in class too, so we can help them figure out the processes they need to do to appropriately incorporate them and consider the ethics around use.”
Heckman said she can’t speak to why employers are choosing to hire students or not, but her experience with industry professionals has indicated that they still deeply value skills like problem solving, critical thinking and teamwork — skills that can’t be replicated with AI. The department’s software engineering class and senior capstone design course focus heavily on these qualities, she said.
NC State’s College of Engineering, where the computer science department resides, is also introducing an “Applied AI” initiative to introduce first and second year students to the ways they can use AI to solve engineering problems. It goes beyond the generative AI and Large Language Models everyone is talking about these days and includes machine learning and classification problems, Heckman said.
Heckman said AI has been disruptive in a sense but, like Pei, she believes it presents an opportunity.
“It shouldn’t ultimately replace us; hopefully it’ll make us more efficient,” she said.
“There’s a lot of things it can provide, but it has to be done thoughtfully and carefully. I hope that our students are going to come out with the awareness of what are the opportunities here, but also where are the challenges and where are the places they need to be careful and thoughtful and respectful of this new technology.”
At Duke, Pei said they are ensuring students know they not only need a new kind of skillset for today’s market to stay competitive, but they must be able to learn in a completely new way.
“The landscapes are changing very fast, but fortunately, these are kids,” Pei said. “They are very young. They are very adaptive. They are very proactive in learning.
“So as long as we provide them the necessary support, work together with them and also learn together with them, that’s the important thing. Because the professors, the faculty themselves, also need to learn.”
For Dockery, while he was aware of the state of the field based on his own observations, he didn’t feel completely prepared by those guiding his education. Little conversation took place within classrooms about the practical state of the field, and advice he did receive was not necessarily sound, he said.
A career development expert from the university came to talk to his class about job searching. The presentation told students not to accept a computer science job that paid less than $90,000 per year, which Dockery said is unrealistic.
“Especially with the market as it is now, it seemed to me that they didn’t know that so many drastic changes had been happening,” he said.
“In terms of actually being in school, I think that NC State adequately prepared me to become a programmer and be able to learn things on my own. But a lot of the things I learned weren’t exactly applicable to what the market is looking for now.”
