Jevin Maltais, a fractional chief technical officer and consultant, has embraced AI to speed up his coding work and tackle unfamiliar programming languages, while acknowledging the need for careful human oversight.Jess Deeks
Jevin Maltais remembers the first time his eyes were opened to the capabilities of generative AI.
It was November 2022, when Mr. Maltais was working as the director of engineering for Humi, a payroll software company. OpenAI had just released ChatGPT version 3.5 and a colleague interviewing coders for a job told Mr. Maltais how candidates were using it to complete coding tests that typically took an hour to do.
“He told me that people were just feeding [the test] into ChatGPT and it was coming back within five minutes,” he says. “I just couldn’t fathom how this could work.”
Mr. Maltais began experimenting – first asking ChatGPT for feedback on his own code. He found the early results unreliable for deeper work. But when version 4.0 launched in March 2023, he saw how AI could be a meaningful tool, particularly for making changes to existing code.
“I won’t go in and manually change the text [anymore],” he says. “I’m asking AI to do it, because it can actually be more thorough or see other areas that might be impacted that I may not have seen.”
The technology is compressing timelines that once stretched for months. Traditionally, implementing a new app feature meant getting a team of engineers and people from sales and business development in a room to discuss what needs to be done, then engineers taking a day or two to plan out how to implement the feature and how long it will take – often a period of multiple months – then returning to the team to present their findings and decide whether to develop the feature.
Now, Mr. Maltais uses AI to build a working prototype in 15 minutes.
“Maybe the version is not very good, but it’s something we can show and play with and that we can continually iterate on, instead of taking months to get to our very first version,” he says.
AI has also allowed Mr. Maltais to work in unfamiliar programming languages. In his current role as a fractional chief technical officer (CTO) and consultant, he recently tackled a product using Kotlin, a language where he lacked deep experience.
“I didn’t have a ton of experience in the languages and the technical stack,” Mr. Maltais says. But he used Claude Code to propose an idea for a new feature. “The things that would have taken me weeks to learn, now I can accomplish in a day or two.”
That acceleration is just one dimension of the shift. AI is also fundamentally changing how developers work together – and whether they need to work together at all.
Jevin Maltais says AI can build early prototypes for apps or new features in as little as 15 minutes. “Maybe the version is not very good, but it's something we can show and play with,” he says.Jess Deeks
Stephen Doxsee, a software developer, CTO and co-founder of SuccessionHR, a software company creating succession planning tools for companies, has found that AI can act as a partner that he can bounce ideas off of.
Mr. Doxsee says that in the past he might have worked with another programmer where one would “navigate” – sharing insights and ideas – while the other would “drive” at the keyboard. Now, AI is that virtual programming partner.
“AI can spit out code and help you cultivate your curiosity by digging deep to understand a problem,” he says. “It’s almost like a super senior developer who knows the ins and outs of lots of different things.”
But the technology has significant limitations.
“It’s not perfect,” Mr. Doxsee says. “AI can be overconfident and it’s not always accurate. You have to check what it does or put guardrails up around what it does so that it doesn’t lead you astray, build the wrong thing or introduce bugs into your code.”
Those concerns are echoed by Jyoti Kunal Shah, an independent researcher and director of application development for ADP, who examines refactoring – the process of restructuring code to make it cleaner or more efficient without changing what it does.
“Providing ... explanations for refactoring decisions can significantly enhance developer trust and transparency,” Ms. Shah writes in her paper, published in July 2024 in The American Journal of Engineering and Technology. “Such explanations connect the AI’s actions to well-known software engineering principles, making it easier for developers to assess and approve the changes.”
She describes a collaborative workflow where AI suggests improvements, explains its reasoning and then waits for a human developer to approve or reject the suggestion. When that happens, Ms. Shah writes, “the developer’s expertise combined with the AI’s speed and pattern recognition can yield optimal outcomes when decisions are transparent.”
Despite AI’s limitations, both Mr. Maltais and Mr. Doxsee see it as an irreversible shift in their industry – one that elevates rather than replaces human coders.
“AI will certainly continue to improve, but I see humans as indispensable to the craft of coding,” Mr. Doxsee says. “We want to deeply leverage the insights AI can provide, but we want to keep people in the driver’s seat.”
Mr. Maltais admits that he feels anxious about AI taking his job. But he says that developers who learn to harness AI will remain in high demand.
“The most powerful thing we can do as engineers is use it regularly and find creative ways to be more productive and provide more value,” he says. “It’s hard to predict where things are going, but finding new ways to use AI is really the main way we can make ourselves more potent as individuals.”