Two JetBrains-supported student hackathons at NYU Abu Dhabi and 42 Abu Dhabi scored teams on whether their software worked and whether they could explain and defend it. The format mirrors a change towards AI coding agents writing much of the code and employers increasingly value junior developers who can verify, test, and take responsibility for it. 

[For more news, click here] 

A confident pitch and a slick slide deck can carry a team a long way at most student hackathons, but the judges at NYU Abu Dhabi on 20 September were not going to be won over that easily. Thirty-five percent of every team's score depended on whether the product worked when it was put in front of them, and each team had to walk the judges through how they had tested it and why they had built it the way they did. 

That kind of rubric reflects where the software industry is heading, one where AI coding tools can turn a problem statement into running code in minutes. In the future of coding, the developers who will stand out will be the ones who can prove their software works and take responsibility for it when it does not. 

The same idea was pushed further two days later at 42 Abu Dhabi, the tuition-free coding school, where Team Kartoshka won with Lattice, an AI agent that sits inside the IDE, keeps track of the other coding agents working on a project, and warns the developer when their work is about to clash. Both hackathons were backed by JetBrains, the company behind IntelliJ IDEA and the Kotlin programming language, and both were shaped around a problem that software employers in the United States and elsewhere are now wrestling with, namely how to judge a young developer when AI can write code on demand. 

Why AI coding Agents are Changing how Developers are Evaluated 

AI coding agents have become part of the daily routine for most professional developers, with the JetBrains Developer Ecosystem Survey 2026 finding that between May and July of this year, 90% used them at work at least weekly and 68% used them every day. Trust in what those tools produce has been far slower to build, and in the Stack Overflow 2025 Developer Survey, 46% of developers said they distrusted the accuracy of AI tools, compared with 33% who trusted it, with their biggest frustration being output that looked right but was not quite there. 

The same shift is hitting the people at the bottom of the profession hardest, according to research highlighted in Stanford's 2026 AI Index, which shows employment among US software developers aged 22 to 25 has fallen by nearly 20% from its late 2022 peak while headcount for their older colleagues has grown. Much of the work that used to train junior engineers, from boilerplate to small bug fixes and simple features is what AI now does which means graduates are expected to arrive already able to weigh trade-offs and take ownership of code they did not write line by line.  

“AI can generate code in seconds. Responsibility for what that code does still sits with the developer,” said Nadia Rinsky, Head of Business Development for the MENA region at JetBrains, who was one of the Hackathon judges. “The students had to go beyond producing code. They needed to understand what they were building, test their ideas, explain their decisions and demonstrate that their solutions worked. AI gives developers more leverage, which makes technical judgement, verification and accountability even more important.” 

A One-Day Build at NYU Abu Dhabi 

NYU Abu Dhabi squeezed its hackathon into a single day on 20 September, organized by the university's Career Development Center with the student clubs HackED and Violet Ventures and powered by JetBrains Academy. The Kotlin Multiplatform Challenge asked teams to build an app that served a community, raised awareness, or tackled a real-world problem, and every prototype had to run on at least two platforms. Beyond the 35% riding on whether the product worked, the rest of the score was split among the strength of the problem and insight, technical execution, and the clarity of the final pitch. 

The top prize went to $ocial, a personalized activity and budgeting app that links to a user's calendar and monthly budget and suggests things to do that fit both their free time and their spending limit, with Waste Track taking second place and Elderly Assist third. A tool that helps companies understand what is involved in expanding into the Middle East and North Africa won the Best Use of JetBrains AI award, while the Judges' Choice Award was shared by two teams that built disaster-management platforms focused on Nepal. JetBrains Academy put up an AED10,000 prize pool, roughly $2,700, along with professional tool licenses for the winners. 

Dana Downey, Associate Dean of Student Affairs & Director, Career Development Center, said, “Our students build on their strong technical knowledge from the classroom by putting it into practice at events like this hackathon. The hackathon allows for innovation under pressure, similar to what they will experience in the workplace. Preparing students for these realities, before they face them on the job, builds confidence and agility our graduates carry into their future beyond NYU Abu Dhabi.” 

Building tools for working developers at 42 Abu Dhabi 

Teams at 42 Abu Dhabi faced a narrower and arguably tougher brief over two days on 22 and 23 September, building a plugin for an IntelliJ-based IDE that met a genuine need in a developer's workflow. They had to account for existing codebases and team practices, and think hard about whether an engineer would want the tool running on their screen all day. 

Lattice took the "Build What Developers Need" prize, with Team Placebo finishing second for an assistant aimed at healthcare development and legacy code, a corner of software where aging systems are hard to read and risky to change. Hybrid Falcons came third with a developer wellness plugin that enforces scheduled activity breaks. 

The format plays to the strengths of the school, which is part of the global 42 network and has no traditional classes or teachers. Students move forward through projects and peer review, so they get used to judging one another's work early on, and that habit is increasingly what employers need from them. 

Fatima Al Foulathi, Acting CEO, 42 Abu Dhabi, stated, “At 42 Abu Dhabi, we are preparing talent not only to adapt to an AI-driven future, but to help shape it. As AI transforms how technology is built, the ability to think critically, solve complex challenges and apply AI responsibly is becoming as important as coding itself. This hackathon gives our students the opportunity to put those capabilities into practice by building, testing and presenting solutions in a real-world environment. Through our peer-to-peer, project-based learning model, we are empowering the next generation of technology talent with the skills, confidence and adaptability to contribute meaningfully to Abu Dhabi’s growing digital and AI ecosystem.” 

What the Abu Dhabi Hackathons Signal for Tech Talent 

Abu Dhabi has its own reasons to care how these students are trained, having committed AED13 billion, about $3.5 billion, to digital transformation and technology adoption under its Government Digital Strategy 2025 to 2027. The emirate expects the program to add more than AED24 billion to GDP and create over 5,000 jobs by 2027. 

JetBrains has a commercial interest here too, having expanded across the Middle East, and student programs have long been a way for tool makers to win over future users. Hackathons have their limits as well, since a prototype built in a day says little about maintaining software for five years or about security and scaling, but the weight these events put on testing, explanation, and accountability matches where the industry is going. 

The students behind Lattice seem to have understood that already, building their winning idea around a future in which several AI agents share one codebase and the most valuable thing a developer can have is a clear picture of what each of them is doing and why. 

Related Articles

UAE Launches National Program for Advanced Sciences and Technologies to Build Homegrown Talent

Exclusive: Nikita Astionov on the Security Risks of Giving AI Agents More Control

Pakistan’s World Context Tests Robot Training Data from Factory Workers at UC Berkeley Hackathon