When Tigran Sloyan was at MIT, it was an embarrassment of riches when it came to recruiters. “Every single tech company [and] non-tech-company showed up two times a year, right to our doorstep,” Sloyan told IT Brew. While he acknowledges that he has math and programming skills, he understands that having MIT on his résumé fast-tracked him to jobs at Google and Oracle.
That’s why Sloyan created CodeSignal, a tool that helps companies identify the right person with the right skills, no matter what their résumé looks like. Today, CodeSignal announced Cosmo, a new chatbot that uses AI to help companies find the right candidate for their technical jobs, partly by determining if the applicant has used generative AI to cheat on their coding tests.
Why CodeSignal? “At the high-level view, we think of ourselves as a skills platform,” says Sloyan. “Skills are going to become more and more central to everything we do. Technology has always created [the need for] new skills and displaced existing ones. We used to ride horses and then all of a sudden, cars came along, and then we didn’t need to learn how to ride horses.”
Companies like Meta, Instacart, and Zoom use CodeSignal with the intention of making the process of hiring technical employees more efficient, effective, and fair, assessing candidates’ coding skills through a structured and standardized process. But skills needed for technical jobs are constantly evolving.
CheatGPT. As tools like GitHub Copilot and ChatGPT explode into the world of coding, it’s critical to understand how the tools function and how engineers can harness them in their work. But that doesn’t necessarily mean companies want potential employees to use them in the application process.
“AI-assisted cheating on technical evaluations and interviews is skyrocketing,” Sloyan said. Before the recent generative AI boom, CodeSignal found that 7–9% of applicants cheated on coding tests. Now that percentage is 20–30%. “Three times as many people are actually using something like ChatGPT to try to find a way to cheat the system,” Sloyan said.
While most companies don’t want applicants to use these tools to cheat, some companies don’t want applicants or employees to use generative AI at all. Sloyan said highly regulated industries like banks will take years to wade through the morass of legal clearance for using generative AI.
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
Current ChatGPT checkers are useless. Sloyan acknowledged that many new tools that claim to spot and root out generative AI use simply don’t work. Most generative AI checkers look at the end result, which is problematic. “AI is incredibly good at producing a result that is almost indistinguishable from what a human would produce,” said Sloyan.
For CodeSignal evaluations, applicants type directly into the CodeSignal tool that Sloyan said can detect what it looks and feels like for an applicant to think through a problem and solve it, versus what it looks like to cut and paste code from ChatGPT.
A new way to measure technical skills of the future. The challenge of finding the right talent and upskilling technical workers isn’t going away anytime soon. But one thing is for sure, what most people are doing in education and training right now isn’t working, including the way Sloyan was hired for his first job out of MIT.
“Humans are just kind of expected to keep up,” said Sloyan. “But none of the infrastructure meant to help humans keep up is changing. You go to college in your twenties and [that] education is supposed to serve you for the rest of your life. People look at your résumé to decide what you know. And if it doesn’t have relevant education or internships [hiring managers think], you’re probably not qualified for this job.”
For Sloyan, hiring technical workers means looking beyond their résumés.
“For the skills that are relevant today,” Sloyan said, “the Stanfords of the world are gonna take a long time to adjust to it, and more and more, the talent that you want and the skill set that you want is going to come from people who take the initiative to learn and teach themselves using all of the online resources available to them.”