For most of school I optimized for the right answer. AvaB began when that stopped being enough — I needed to know whether the student on the other side of the screen actually understood. As team leader of five, I built the platform and its personalization layer, routing each learner by WHY they got a problem wrong, not just whether. Our first rules were too aggressive: they pushed quick students ahead and quietly lost the ones who needed a slower path. That failure taught me what no olympiad had — a model that is right on average can still fail the one student in front of you.

Today avab.vn is live — around 1,500 learners and parents, roughly 230 paying enrolments a year, built and run by high-schoolers, and recognized from our school's contest up to the national student-startup round. But the result I care about most is quieter: a student I'll never meet opens it on an ordinary Tuesday night and finds a lesson that meets them where they are. That is why I now build for the learners technology skips. I want the kind of US program where MIT's make-real-things culture and Princeton's service of humanity are the same instinct — where going deeper into CS and AI and putting a model responsibly in front of a child are one ambition, not two.

And I don't only build in software. With a VEX Robotics team, I helped design and iterate a machine that earned our way to the World Championship — held here in the US. It's the same instinct as everything else on this page: design it, watch it break, fix it with your hands, and go again.