Hackathon in the AI Age

On September 11, 2026, 44 students from the Master of Financial Technology and the Master of AI & Data Science programmes at Frankfurt School sat down with guests from OpenAI for a hackathon. My first reaction was skepticism. from OpenAI for a hackathon. My first reaction was skepticism.

Ninety Minutes? Really?

I still remember hackathons that stretched across entire weekends. Hackathons were meant to be all-nighters fueled by energy drinks. So, when I heard we'd have just 90 minutes, with a team that barely knew each other, I assumed we'd end up with a half-baked slide and …apologies?

Our team was formed on the spot — Adrian, Yihan, Praavar, Berzan, and me — was a mix of three different programme intakes, two master's programmes, and four nationalities sitting around one table. We'd exchanged maybe a handful of words before the count-down started.

What we walked away with, 90 minutes later, was a full pitch deck, a working demo, an in-depth business model analysis, and a promising start-up idea. And we weren't the exception. All nine teams in the room pulled off the start-up plan, each with a strikingly different idea. Some teams-built AI assistant for renting apartments; others designed apps to help students keep pace with fast-talking professors’ mid-lecture. I am amazed by the range of creativity in the room.

The Idea was the Hard Part

It started slowly. Writing the code, mocking up the demo, designing the deck: none of that was the bottleneck. The most time-consuming part of the whole sprint was finding a truly good idea and shaping it into a story worth telling.

We spent the first valuable minutes just thinking and talking, examining our own lives as students in Frankfurt for a problem worth solving. And then it clicked. Adrian, it turns out, owns four microwaves. I have a coffee machine gathering dust in my kitchen. Neither of us wanted the hassle of listing them on some clunky secondhand marketplace, haggling with strangers over a few euros.

So, we asked: what if it worked more like swiping? A dating app, but for your things. You list what you have, you list what you want, the app finds your match, and you barter them. We named it on the spot: Swable. You swap things, like Bumble!

From Idea to Pitch, Powered by ChatGPT

Everything moved fast after we had the concept. Really fast. With ChatGPT as a helpful teammate, we coded the MVP, sketched our business model, and built out the pitch deck during the next hour.

When we finally stood up to pitch, the room responded better than we'd hoped. People laughed at the microwave story, nodded at the swapping concept, and leaned in for the demo. We walked away with 5 original OpenAI Codex hoodies.

The Rewards

There was something even nicer than the hoodies: the feeling of building something real, start to finish, with people I'd just met, and having it resonate with a room of strangers. That's a rare kind of validation, and it's the kind that sticks with you for a long time. But the story didn't end with the experience and the hoodies. When it was all over, the five of us looked at each other and agreed on the same thing: we don't want this to end here.

We're taking Swable out of the hackathon room and into the real world. We believe there's a better way to help people pass along things, and we want to build it for real. We are now continuing to work on Swable to see where the idea goes next. Stay tuned!