Kris Jordan uses AI in software engineering

When Kris Jordan set out to integrate generative artificial intelligence into his COMP 423: Foundations of Software Engineering course, he wasn’t just experimenting with new tools. The professor of the practice in computer science was preparing students for the future of work.

Thanks to support from Carolina’s AI acceleration program, Jordan distributed OpenAI application programming interface keys to all 200 students in his spring 2025 course, giving them direct access to cutting-edge language models like OpenAI’s GPT-4.

An API enables students’ programs to integrate directly with state-of-the-art AI models via a request and response protocol called an API call. The API key unlocked and securely authenticated each student’s access to the API.

“This was the first time most students had ever made third-party API calls, let alone integrated with a large language model,” Jordan said. “It opened up a whole world of possibilities.”

Building with AI

Students didn’t just learn how to use AI — they learned how to build with it. Every team in the course was required to incorporate Azure’s OpenAI services into their final project. Results ranged from a virtual academic adviser that could recommend courses based on student goals to a weekly office hours summarizer that helped instructors identify common student challenges.

One team built a job prep assistant that simulated technical interviews and evaluated responses, while another created a chatbot to help students find campus clubs that matched their interests.

Students learned the fundamentals of prompt and context engineering, explored latency and user experience challenges and grappled with the limitations of generative models. They reported using their AI skills at companies like SAS, working on projects involving large language models.

“They felt confident and prepared. That’s a huge win for brand Carolina,” Jordan said.

The AI acceleration program provides opportunities and resources for Carolina community members to explore and advance AI technologies. University faculty, staff and students who seek cloud-based computing resources to pursue their AI initiatives may apply to the program.

Aligned with industry best practices

To guide responsible use, Jordan introduced a benchmark: no more than 25% of a student’s code should be AI-generated, mirroring industry practices at companies like Google. “It wasn’t something we enforced punitively,” he said, “but it helped students think critically about how much they were relying on these tools.”

The success of the course was made possible by strategic collaboration, with the program providing the funding, and ITS delivering the platform that enabled access to Microsoft Azure’s OpenAI services. Computer science students deployed their projects using Carolina CloudApps, a container-based development environment run by ITS, which gave them hands-on experience with modern software engineering practices.

“We couldn’t have done this without AIAP,” Jordan said. “It’s easy to look at a number like 200 students and think it’s small, but this is half of a graduating class of majors in a single semester. Thanks to this program, every CS student at UNC is gaining real experience building with AI.”

Technological turning point

Jordan also praised the infrastructure support from the CloudApps team, noting that students were able to deploy full-stack applications and learn production-level engineering skills in a real-world environment.

Jordan is now teaching in Cape Town, South Africa, with a group of Carolina students, and using the semester to retool his course for the spring. He hopes to shift toward an “AI-first” curriculum that explores agentic systems, asynchronous processing and real-world deployment challenges.

“This is a technological turning point,” Jordan said. “We’re not waiting to see how it plays out — we’re helping students lead the way.”