Recently, I’ve tackled a new and interesting endeavor in my career - volunteering as a hackathon judge. I always found the hackathon culture inspiring and exciting, and not without a reason - it is a place of creativity, unique fresh angles, and just general fun vibes. Now, obviously, with the AI revolution, there’s been a significant shift in a lot of sectors of our lives, and for better or worse, the space of competitive programming is no exception. There are things that have become easier to do for competitors, but because things are easier to implement, it may bring out an unexpected twist to the process that I think is important to talk about. In this article, I will be exploring things that changed with the introduction of AI/Vibe coding tools, overviewing the common pitfalls that young engineers face, and I will try to provide a perspective from the judge's seat, so that hopefully, it may help any potential reader stand out in any future hackathons they participate. I should note that, personally, I think that due to the innovative spirit and nature of such competition, I may come across as very critical/conservative towards the use of AI in certain parts of the competition. Further, since I find many such competitions attract a lot of young and less experienced engineers, I think that they should also be viewed as a powerful learning platform - a place where young, talented people shine - another reason why I think AI tools may be harmful when overused, but more on that later. a place where young, talented people shine Just to be clear, I am NOT against the use of such tools; in fact, I think with correct usage, such tools will, on the contrary, allow you to exceed above and beyond and actually benefit your career growth. That being said, they are still tools, and to delve deeper into the topic, let’s first explore the structure of a hackathon. NOT tools, Structure and Phases of a Typical Hackathon Personally, I think there are 3 important parts of any hackathon we can explore: Brainstorm Phase This is where the ideas are born, refined, and discarded. Team members come together, mentally wrestle the problem at hand, and set the direction that they will decide to go, sometimes correctly, sometimes not. Ideas are born, refined, and discarded. For all intents and purposes, this can be considered a foundation step, and is the perfect place for creativity to shine. Implementation Phase At the point when the team agrees on the problem and its solution, it’s time to actually implement it and turn the idea into something tangible. With modern tooling, this part is significantly sped up and is arguably the phase where AI tools shine the most. You can scaffold projects faster than ever, wire APIs in minutes, generate UI layouts almost instantly, and in some cases, even produce entire working demos with surprisingly little manual effort. From a purely execution standpoint, this is incredibly powerful. However, this is also where I’ve seen one of the biggest shifts compared to earlier hackathons. Because implementation has become easier, teams often end up building more than they can fully understand or explain. Features pile up quickly, abstractions stack on top of abstractions, and at some point, the team itself starts losing a clear mental model of what is actually happening under the hood. From a judge’s perspective, this becomes very noticeable. building more than they can fully understand or explain Demo/Presentation Phase Demo/Presentation Phase Finally comes the demo and presentation phase - the only part of the hackathon that judges actually experience directly. No matter how impressive the internal architecture is, or how many hours were spent coding, this phase defines how your project is perceived. And this is where many strong projects, unfortunately, fall short. Interestingly, AI can help here as well - but again, only if used correctly. AI can assist with structuring a presentation, refining language, or simplifying explanations. What it cannot do is replace genuine understanding. If the team doesn’t fully grasp the solution, no amount of polished slides will hide that. How to Stand Out Amongst the Competition There are a couple of key points to understand about the modern hackathon projects - almost 90% of them have some form of AI influence, and to actually stand out among the competition, leave a good impression on a judge and benefit most from the learning effect a hackathon has on an engineer, most of it comes down to intentionally limiting the use of AI during the most innovative parts of the hackathon. Here are some key bullet points: 1. Actually brainstorm the idea yourself - do not delegate the innovation to AI. The human brain is still and probably will be for a long time the best source of innovation humanity has at its disposal. As such, do not delegate the actual brainstorming to LLMs. Due to their nature, the ideas they generate have a pattern to them, and a judge can feel it. I’ve lost count of how many projects I opened, and it was the same old formula - an idea that sounds awesome but is actually a bland variation of the same pattern. For example, say you decide you want to tackle a HealthTech problem; a human idea would feel relatable, a problem that the people implementing it may actually face: a reminder app to apply face care products or a skin weather app that predicts acne flares. If you prompt AI to brainstorm the same solution, it would probably generate something like a MentalCare app that helps users stay connected with their wellness, tracks bad moods, yada yada yada. Do not get me wrong, mental health is a super important domain and has a sufficient space for innovation - but my point is that AI-generated ideas will feel AI-ish. And as a judge, I think that this is not a place for such a tool. Let your imagination fly - think of problems you and your friends face, Google domain discussions online, and think outside the box. Such an effort shows in the final results, and even if you don’t win, it will teach you to think like an innovator. 2. Implement the app by limiting vibe coding to templates and boilerplate tasks. Earlier, I mentioned that during the implementation phase, AI tools speed up the process significantly; however, it is important to use them wisely. I also mentioned that for a young engineer, hackathons are a place of learning. The topic of AI influence on junior devs is an important topic that merits a blog post on its own, but to sum up, it is crucial to understand the following point: AI should not code the whole of your project if your goal is to learn something. Ask AI to generate a conventional project structure, outline the class names, but do put in the effort to actually implement the things you generate at least a couple of times. While it will not yield results as fast as opposed to your competition, be the Tortoise in this race. You will gain valuable knowledge that will help you in future projects and real jobs. This part won’t help you win the hackathon; it is usually invisible to the judge, nor does it affect the resulting solution. However, I still think that you should be mindful of the effect of AI overuse. 3. Prefer to design the app with individuality. If applicable - prefer to design your app individually. Modern AI platforms such as Lovable do a super impressive job at generating a website in one prompt. It will look super professional, polished, but also - no surprise - look AI-generated. Being too polished may actually hurt the overall feel of authenticity. Now, I understand that not everyone is a UI/UX designer nor wants to dabble with FrontEnd and design, and I totally resonate with it - I prefer backend work as well. However, having an individual design may actually help score “respect points” with the judge, as it will feel like more effort went into the design than just a simple prompt; as such, this can ultimately help you stand out. 4. Present the app using your own voice - ideally your own face. The presentation phase is an important part of the hackathon judging process - impressions are important, and the judge sees the demo video first. As such, I will not delve into the tips and tricks of the actual demo, such as the hook, content, etc. But I will mention that if you present the app using your own voice and a camera, the judge will look at it more favorably than, say, for example, a competitor that generated the voiceover with AI. Again, authenticity, human feel, and effort are the new gold standard in hackathons. This is especially a good experience for you if you are shy and find the idea of speaking on camera stressful. I am an introverted person myself, and I’m still collecting the courage to start public speaking at conferences, but there is no shame in a person who may be a bit less confident during the presentation. For the longest time, I was insecure about my English accent - I was born and raised in Moldova, with a mix of Russian and Romanian speaking. I obviously have a noticeable accent, but first of all, most people don’t care if you articulate words clearly, and if you are hard to understand or don’t have fluent English, even better, you will receive valuable practice! I had a submission with a gentleman who had a very thick accent that maybe was a little harder to understand, but he used AI-generated captions, and I was super impressed. He got the practice in, and I understood everything. This is a perfect example of putting in effort - as opposed to delegating everything to AI. Conclusion Hackathons are, at their core, not about winning prizes or shipping production-ready products. They are about learning, experimenting, and pushing yourself into slightly uncomfortable territory — technically, creatively, and sometimes even socially. AI has undeniably changed the way hackathons work. It has lowered the barrier to entry, accelerated implementation, and made it possible to build things that would have been unrealistic just a few years ago. That is a good thing. But as with any powerful tool, its value depends entirely on how intentionally it is used. From the judge’s seat, what stands out today is not who used AI the most, but who used it with purpose. The teams that leave a lasting impression are the ones that still think for themselves, still make conscious trade-offs, and still put something personal into their work — whether that’s the idea, the design, the implementation, or the presentation. with purpose If there is one takeaway I’d like to leave you with, it’s this: don’t outsource the most human parts of a hackathon to machines don’t outsource the most human parts of a hackathon to machines Let AI help you move faster where speed matters, but keep ownership over the parts that shape your thinking, your learning, and your identity as an engineer. Even if you don’t win, you’ll walk away with something far more valuable - sharper intuition, stronger confidence, and a clearer understanding of how you build and present ideas. And from my experience, those are the people who eventually stand out far beyond hackathons. That, in my opinion, is the real win.

From the Judge’s Seat: Standing Out in Hackathons in the Age of AI

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Apache Beam on GCP: How Distributed Data Pipelines Actually Work (for REST API Engineers)

The Noonification: Use This 7-Step McKinsey Framework to Solve Any Problem (1/10/2023)

The Noonification: A Taxonomy of Inclusiveness (1/11/2024)

The Noonification: What is the InfiniteNature-Zero AI Model? (11/19/2022)

10 Ways AI Has Changed Our Lives

100 Days of AI, Day 8: Experimenting With Microsoft's Semantic Kernel Using GPT-4

Apache Beam on GCP: How Distributed Data Pipelines Actually Work (for REST API Engineers)

The Noonification: Use This 7-Step McKinsey Framework to Solve Any Problem (1/10/2023)

The Noonification: A Taxonomy of Inclusiveness (1/11/2024)

The Noonification: What is the InfiniteNature-Zero AI Model? (11/19/2022)

10 Ways AI Has Changed Our Lives

100 Days of AI, Day 8: Experimenting With Microsoft's Semantic Kernel Using GPT-4

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps