Guides
Hackathon ideas: how to come up with a project that the jury will love
Where to get ideas, which projects win more often than others, and why you should ruthlessly cut down the scope of your concept.
“We don't have an idea” is the most common roadblock at the start of a hackathon, and also the most overrated. It's not the most original idea that wins, but the one that is best assembled and demonstrated. Still, you have to start somewhere, so let's look at where ideas that impress the jury come from, and how not to drown in your own concept.
Where to get ideas
The best projects don't grow from “let's invent something innovative,” but from a specific frustration.
- Your own pain point. What annoys you personally in your studies, work, or daily life. Most likely, you are not the only one annoyed by it.
- Tracks and prizes from the organizers. The hackathon page on Stavleak lists the tracks and what prizes are offered for them; read them literally: they state directly what the jury wants to see.
- Real processes around you. Queues, paperwork, routine tasks that are still done manually.
- A fresh technology paired with an old task. Not “let's build an AI,” but “let's eliminate one specific manual action.”
Don't chase after trying to amaze the world. A problem that the jury understands in ten seconds wins over a brilliant but incomprehensible one.
A separate note on AI ideas
Nowadays, you can build a wrapper around a chat model in an hour, which is exactly why wrappers lose: they are brought in by the dozens, they look identical, and they are impossible to remember. What wins is not “we added AI,” but a specific task where AI solves a real pain point, works with its own data, and does something a bare model cannot. We discussed how to distinguish a real project from a wrapper separately.
Which projects win more often
If you look at the winners of various hackathons, they have something in common. A clear problem that everyone in the room can relate to. A working demo, not just a concept. A narrow but fully completed idea. And a story that is easy to retell in a single sentence.
Usually, the opposite things lose: a scope that is too broad to assemble in time, a solution to a problem that nobody has, and a project that sounds beautiful in words but doesn't work on the screen.
Ruthlessly cut down the scope
The most common mistake of beginners is aiming for a “platform,” “ecosystem,” or “super-app.” You can't build that over a weekend, and you'll end up presenting a half-baked dream. Take one feature, make it work, and show it. One live scenario beats ten mockups.
- Formulate your idea in a single sentence using the template: “we help someone do something faster or easier.”
- Cross out everything that does not affect this single scenario.
- Check if you can show the result in two minutes. If not, cut it down further.
What to do if you have absolutely no ideas
Come to the venue and listen to what others are discussing and what the organizers say at the opening. Ideas spark off one another in conversation, rather than being born in silence in front of a blank screen. Looking at others' work also helps: browse past and current events in the hackathon catalog, look at their tracks and tasks to understand the level and find your own angle.
An idea is just the start. From there, everything is decided by the team, the 48-hour schedule, and the pitch, and we have separate breakdowns about those.