A builder who turned a familiar market-day problem — where does the used oil go? — into an AI-powered recycling system, now a global category winner at the MIT Global Appathon 2026.
18 Court Market has been part of my life since childhood, next to my house, where I'd trail after my mother on every visit. And on every visit, I watched the same routine: vendors tipping used cooking oil straight into the drain after each dish.
For years, I didn't see it. Not because it was hidden, but because I'd stopped looking. It was just the market, the way it had always been, and I'd long since filed it under things that didn't need explaining.
Then I came across a video of oil recycling efforts in China, and a question I'd never thought to ask finally surfaced: why not here, in a market I thought I already knew? So I did the thing I hadn't done in a decade of visits. I asked the vendors directly why they weren't recycling their oil.
Almost every answer was the same: no one had ever asked them that before.
I had assumed, without realizing I was assuming anything, that the oil in the drain meant no one cared. What I found instead was a gap I'd mistaken for indifference, not a lack of concern, but the absence of anyone who'd bothered to ask, or to offer an alternative. The market hadn't changed in all the years I'd walked through it. I had just finally learned to notice what I'd been overlooking the entire time.
That realization became the foundation for EcoBin, not a problem assigned in a classroom, but one I'd been walking past my whole life, waiting for someone to finally ask.
"We kept hearing the same story at market after market — vendors pouring small amounts of used oil down the drain because there was nowhere convenient to take it. EcoBin exists to make the right choice the easy choice, one bin at a time."
Interviewed local food vendors, small restaurants, and households at markets across Bangkok and Nonthaburi to understand how used cooking oil is really being handled.
Curated a 1,162-image dataset, trained a YOLOv11-cls classification model to 99.64% accuracy, and built a working prototype around a Raspberry Pi 5, camera, and water pump.
Advanced through three rounds with NSTDA, Thailand's national science and technology agency, and received a $300 development grant to keep building.
Presented EcoBin at CMKL University's AI Innovation Summit and Sustainability Expo, alongside a working app and LINE Official Account, and received a scholarship offer through mentor recognition.
Won the Youth Team Category against 2,146 teams from 141 countries and presented EcoBin at the MIT App Inventor Global Education Summit.
Deploying pilot bins to collect real-world data, retraining the model with active learning, then rolling out EcoBin v2 more broadly.
Whether you're a market, a recycling partner, or just curious about EcoBin — try the app or reach out to the team.