Transitioning from competitor to Host Scientific Committee member for IOAI 2025 offered me a perspective few get to experience. While I brought fresh insights from my silver medal performance in 2024, what I didn't expect was how this role would completely transform my understanding of what olympiads truly accomplish. As a young committee member working alongside experienced AI
educators and researchers, I found myself in a unique position—both contributing to and learning from the evolution of this new olympiad. The technical challenges of problem design initially captured my attention, but it was the deeper purpose of our work that ultimately fascinated me most.
Casual Conversations That Changed My Perspective
What truly transformed my understanding of IOAI's purpose happened duringlunch breaks and casual conversations with team leaders from around the world.
I remember chatting with Antônio, a team leader from Brazil, whilewaiting for coffee during one of our breaks. When I asked about his team's preparation process, his answer surprised me.
"For us, just participating means we've already won," he explained. "Before IOAI announced its first competition, AI education was barely on our educational ministry's radar. Now we have universities offering AI preparation courses for high school students."
Similar conversations followed. Team leaders from GAITE mentioned howtheir selection process had spurred the creation of weekend coding camps in regions previously without programming education.
These fundamentally shifted how I viewed our work on the committee.
Late-Night Reflections
These conversations stayed with me, especially late at night when I wasworking on problem drafts in my dorm room. I began questioning my approach to designing problems.
I had been focused on creating technically impressive challenges thatwould showcase cutting-edge AI techniques—problems that would have impressed my professors and peers back home at China. But what good was a sophisticated reinforcement learning problem if it required computing resources that most countries couldn't access?
I started experimenting with different problem frameworks. Could we designchallenges that scaled across different resource levels? What about problems that highlighted AI applications relevant to developing regions?
One night, after reworking a computer vision problem to require less computational power, I messaged Jiayu, another student member from Singapore who had become a friend during the competition.
"Do you ever think about how these problems might be used beyond the olympiad?" I asked.
His response was immediate: "All the time. My teacher back home usesmodified versions of last year's problems for our AI club. Most students can't solve them completely, but they learn so much from trying."
That simple exchange confirmed I wasn't alone in seeing the broadereducational potential of our work.
Learning From Team Leaders
The most valuable insights came from observing how experienced teamleaders approached problem design. During one review session, Dr. Sharif, a veteran of multiple olympiads, spent nearly an hour questioning whether a particular problem was testing the right skills.
"The best olympiad problems aren't about knowing the latest algorithms," he explained to us newer members. "They're about testing how students think about problems—how they decompose complex challenges
and apply fundamental principles creatively."
Another team leader added: "Remember that for many countries, theproblems we create here will become teaching materials for the next generation. They need to be more than just difficult—they need to be illuminating."
These weren't revolutionary statements, but hearing them from people whohad dedicated decades to olympiads made them resonate differently. I began seeing our problems not just as evaluation tools but as educational resources that would live far beyond the competition itself.
Small Contributions, Bigger Picture
As a junior committee member, these insights did change how I approached my
contributions.
When reviewing problem proposals, I started asking different questions:Would this problem work in resource-constrained environments? Does it showcase AI applications with real-world relevance? Could it be adapted into educational materials for different levels?
During one feedback session back at IOAI, I suggested adding supplementary educational materials to accompany our problems—simplified versions that could be used in classrooms with limited resources. To my surprise, several senior members expressed enthusiasm for the idea, with one mentioning they had been
considering something similar.
I won't pretend this was my unique insight or that I transformed thecompetition overnight. But it felt meaningful to contribute even a small perspective to the ongoing evolution of IOAI.
Personal Growth
Perhaps the most significant change has been in my own perspective. Beforejoining the committee, I viewed olympiads primarily as talent identification systems—ways to find and recognize exceptional students. And they certainly serve that purpose well.
But I now see them as something more powerful: catalysts that can accelerateeducational change and democratize access to cutting-edge fields. When a country decides to participate in IOAI, they're not just sending a few talented students to compete—they're often initiating broader investments in AI
education that benefit thousands of students who will never attend an international competition.
This perspective has made my work on the committee feel more meaningful.We're not just creating interesting technical challenges—we're helping shape how the next generation around the world encounters artificial intelligence.
As we continue preparing for IOAI 2025, I carry these insights with me.And while I'm still one of the less experienced voices in the room, I've found that bringing this educational perspective to discussions has value. After all, olympiads should be about more than identifying who's best—they should help everyone get better.
Next time, I'll share thoughts on what makes an excellent AI olympiad problem, drawing from both my competitive experience and what I've learned from the brilliant problem designers I've been fortunate to work alongside.