To make the most of this, I’ve spent the past six months joining different hackathons organised by Gradio, Kaggle, Gemini, OpenAI, and more.
I’ve had all kinds of outcomes: top-3 finishes in a few, consolation prizes in others, some results still pending, a few where I didn’t even make the cut, and a couple where I couldn’t finish in time to submit. But irrespective of the result, each hackathon has been a great learning experience and in this article, I share some of my reflections and takeaways from the journey, in no particular order.
1. Every idea starts small
An idea is fragile in the beginning. It has to be nurtured and needs time to grow. In the beginning ideas may appear messy, too wide and probably too simple. However as you spend more time, iterate on them, you’ll see them slowly converging to something meaningful and workable.
I always thought that there is an a-ha! moment and you get ideas via some inspiration. In reality, it takes a lot of effort to go from a basic form to something that can be worked upon and that is the beauty of ideation. More often than not, the final product looks a lot different than the first sketch and honestly, I enjoy this ideation phase even more than the submission, because that’s when creativity comes alive.
This phase reminds me of the Double Diamond model from the Design Council which shows two stages of the design process:
- The first diamond: Discover → Define (understanding the problem)
- The second diamond: Develop → Deliver (creating and refining solutions)
2. The 80/20 Rule
In the excitement of submission, it’s easy to get straight into coding. I’ve done this myself. Once, I rushed ahead with an idea that felt exciting, only to realize halfway that it wasn’t realistic and didn’t even fit the theme. Since it was a weekend hackathon, I had already lost crucial time.
What I’ve learned is simple: spend 80% of your time upfront thinking and validating your idea, so the build takes only 20%. This saves you from a lot of wasted effort and painful rework.
3. Keep It Simple, Don’t Overcomplicate
As Dieter Rams, the visionary designer, famously said,
Adding more features to your app can be great, and useful ones can make your product stand out, but sometimes we go too far, piling on too much, and the product ends up distracting the user instead of helping them. There is also a term for this called Feature Bloat.
Take, for instance, the traditional TV remote. It has so many buttons and we don’t even use half of them. I mean, why include stuff that is not usable and takes up so much space? Instead, keeping only the most-used and essential buttons can reduce clutter and make the remote far more useful.
The tweet below nicely encapsulates my own thoughts in a very subtle manner. While its easier to create features today, thanks to AI assistance, make sure you use that help to build something meaningful instead of just slop.
4. Don’t neglect the code
Putting together code today is easier than ever. With coding agents, you can spin up a working app quickly, but writing code that’s clean, readable, and safe is still a skill. If you’re aiming to turn your hackathon project into something production-ready, you need to review your code yourself. I’ve seen people hardcode API keys or ignore basic security practices just to make it work. That’s fine for a quick demo, but dangerous beyond it as can be seen in the viral example below.
Matt Palmer from Replit has written an excellent blog addressing this very problem. In his write-up on Secure Vibe Coding, he has captured the security issues that come with modern vibe coding. It’s definitely worth a read.
Matt shares a practical security checklist and several best practices to help keep hackathon projects safe. For instance, he emphasizes the need to keep dependencies updated, since most vulnerabilities arise from outdated libraries. He also stresses proper error handling to avoid exposing sensitive details in error messages. Cookies should be secured and for file uploads, he recommends validating file types and sizes and scanning for any malicious content. Lastly, he advises implementing rate limiting on all API endpoints, especially those related to authentication. The figure below is adapted from his blog itself.
5. Think Like a Product Person
Most hackathons today don’t just ask for a working prototype. They also expect a project write-up and a short demo video that looks professional. This means you need to focus on every aspect of your project, from ideation to final presentation, much like a product person. Coding is important, but it’s not the only consideration. Your use case and how your project delivers value matter just as much. You’re not only defining the problem but also thinking about the user journey and how your product solves that specific need.
For instance, the Gemma 3n Impact Challenge, which went live on Kaggle earlier this year, required participants to include a video demo, a technical write-up, a public repository with the code, and a live demo of the hosted project as part of their submission. And trust me, creating professional-quality videos isn’t easy.
That’s why it’s essential to plan time for presentation. I usually record videos, edit them, add captions, and even do a voice-over. It takes effort, but if you’ve worked hard on your idea, showing it well is just as important as building it. Here’s my team’s attempt at creating the video presentation for the Gemma 3n Impact Challenge.
6. Look around for inspiration
The best hackathon projects solve real problems. Start by observing actual pain points around you and ask yourself what feature you had always wanted that didn’t exist yet.
In these hackathons, your first users are the judges, so it helps to think from their perspective. For instance, I participated in this year’s Meta Kaggle Hackathon, where Kaggle provided 15 years of data for the community to analyze and uncover trends and insights that could be useful. For this challenge, I approached the data from Kaggle’s point of view and tried to understand what signals matter most for their community growth and whether Kaggle’s recent efforts were helping revive engagement.
My submission, which earned second prize, focused on why many community members go silent after joining and what can be done to reactivate them? I believe this addresses a real pain point not only for Kaggle but for many other tech communities, and I’m glad it resonated with the judges.
Social media also serves as a great source of inspiration. On platforms like X or Reddit, people are very vocal about what frustrates them. While there is plenty of noise, if we are able to sift the signals, we can have ideas that could be turned into impactful solutions. When you focus on solving a real, specific problem, the solution often ends up helping a much wider audience.
7. Create an Idea Bank
The best way to have a good idea is to have lots of ideas.
— Linus Pauling
Ideas don’t wait for hackathons and inspirations can strike anytime. Therefore Instead of letting them slip away, create a simple idea bank to capture them. It doesn’t have to be fancy. A simple notes app, a Notion page or even a spreadsheet works equally well. The point is to store every spark so that when a hackathon comes around, you’re not starting from zero.
It’s also important to regularly organize your ideas. As David Allen emphasizes in his writing on capturing good ideas, it’s not enough to just have the tools; you also need the habit of organizing them later. He notes, “If you leave emails unprocessed in your inbox, or paper notes piled up in a briefcase or notepad somewhere, the whole process is defeated, and your motivation to continue will disappear.”
By building this habit, you’ll already have a pool of ideas ready to refine, adapt, and build on.
8. Choosing the Right Hackathons to Join
While hackathons are fun, it’s just as important to know which ones to skip. In the beginning, I signed up for almost every event I came across. At one point, I ended up doing four back-to-back hackathons, two that lasted two months each and two weekend sprints. It was exciting, but exhausting, and I burned out quickly.
Now I’m more selective about where I invest my time and energy. Picking the right hackathons means you can focus better, build stronger projects, and actually enjoy the process instead of running yourself into the ground.
9. Share your work with the world
Creating a working app takes real effort, so whenever you can, share it with the community and there’s no better place than X to do that. Showcase your project to the world, whether it wins or not. You never know who might find it useful or whose problem it might solve. Sharing openly also helps you get feedback so that you can continue to improve upon your idea. Who knows, you might even end up teaming up with someone for your next hackathon.
Final Thoughts and Finding Your Next Hackathon
Participating in these GenAI hackathons reminds me of machine learning competitions on Kaggle. When you first start, everything feels overwhelming, but gradually you develop a set of best practices and a reliable toolkit you can fall back on and adapt quickly. Over time, I’ve built a personal toolbox that brings together building blocks and lessons from every competition I’ve taken part in.
So where to look for new hackathon announcements? The best place is undoubtedly X and following the right people and communities can help you easily find them. Other than that, platforms like Kaggle, Devpost and dev.to are also some great places to look for upcoming events.
If you’ve made it this far, I’m genuinely glad you did. Hope you got your daily dose of motivation and are ready for some hacking. All the best!
Image Credits
Featured image: The Double Diamond by the Design Council — licensed under CC BY 4.0.
