Qualifiers for the LaunchPad Challenge are over. Congratulations to the Top 20 teams moving on to the Builder Showcase.
Every team that submitted put time into making something work, explaining why it matters, and getting it in front of judges. That is difficult work. The Top 20 earned their place, and we are excited to see them present at NUS on 18 August.
This challenge came from an experience I had in November 2025.
I attended the Cursor Hackathon at SUTD, led by the team at 65 Labs. Before I could even sign up, Agrim Singh, one of the organisers, asked me a series of hard questions about my real-estate idea.
At the time, I thought I needed a convincing idea. Agrim’s questions made me think about the work more seriously. What problem was I really solving? Why did AI need to be part of the solution? Could I show what the system actually did, rather than just present a polished interface?
That hackathon introduced me to many strong builders, some of whom are still my friends today. It also changed the standard I held for projects: a good demo matters, but it should be backed by clear reasoning and a system that can hold up.
Since then, hackathons and AI builder meetups have grown quickly. That growth has brought more people into the space, which is a good thing. But the format can become repetitive. When an event is compressed into a day or two, teams naturally optimise for what can be understood in a few minutes. The result can become more about presentation than the quality of the system underneath it.
There is nothing wrong with strong design or a clear demo. Both are important. But an AI project should also show that the team understands the problem, has made sensible technical decisions, and can be honest about where the solution does and does not work.
That is why we built the LaunchPad rubric around three simple questions:
- Do you understand why the problem is hard?
- Does the solution work, or come honestly close?
- Can you explain the decisions you made and why?
We wanted the LaunchPad Challenge to give students a place to put forward their best work, learn what production-minded AI building demands, and meet people who can recognise potential beyond a flashy demo.
The last part matters a lot to us. A team can spend weeks working through a difficult problem and still miss out on a prize. Their work should not become invisible because they did not have the slickest presentation. We want students to leave with useful feedback and connections to people who care about what they are building.
The LaunchPad Challenge is not perfect, but the team is trying to give students the strongest opportunity we can while we are at the National University of Singapore. Excellence is not only a grade on a transcript. It is also the number of students we can support through the work we do.
If you are an industry leader, recruiter, founder, or technical practitioner, we would love to have you at the symposium. You can register your interest to attend, mentor students, or judge the Top 20 finals through our industry registration form.
Finally, thank you to OpenAI, Zo Computer, Griffin Labs, Agnes AI, and GMI Cloud for making this possible. Your support gives student builders more room to build, learn, and meet the people who can help them take the next step.