My First Hackathon

Walking into UofTHacks for the first time felt overwhelming in the best way possible. The energy in the room was unreal: hundreds of students, laptops open, whiteboards covered in ideas, and conversations about APIs, models, and product pitches happening everywhere. What surprised me most was not just the scale, but how open everyone was. Within the first hour, I had already met students from different universities, backgrounds, and skill levels, all excited to build something ambitious. It did not feel competitive in a negative way. It felt collaborative.

Of course, hackathons are not just about coding. There was great food flowing throughout the weekend, and honestly, those meal breaks became some of the best networking moments. Sitting with people you just met, talking about projects, internships, or random tech debates at midnight made the experience feel bigger than just one build. It felt like joining a community.

That said, sleep was optional. At some point in the night, exhaustion hit. Finding a proper place to sleep was harder than expected. Every couch and quiet corner was taken. Eventually, I gave up and tried to get some rest on a chair. It was not comfortable, and I do not think I slept more than an hour or two, but weirdly, it felt like part of the authentic hackathon experience. Waking up slightly disoriented, grabbing coffee, and jumping straight back into debugging real-time event flows was chaotic, but memorable.

Our project, Amplify, came out of that environment. We wanted to build an AI-powered behavioral intelligence layer for e-commerce storefronts. Instead of treating every visitor the same, Amplify profiled behavioral traits in real time and adapted the storefront dynamically based on inferred intent. Using Next.js, React, and a TypeScript event bus, we processed live shopper interactions in roughly 200 milliseconds. We built inference logic that triggered UI changes once confidence thresholds were reached, and we prioritized explainability by visualizing behavioral insights so users could see why the interface was adapting.

The hardest part was not writing the code. It was managing scope under pressure. At 2 or 3 a.m., debugging real-time event pipelines while running on almost no sleep forces you to think clearly and architect efficiently. We had to modularize quickly, define clean data flows, and ensure that our system was not just impressive, but defensible.

By the end of the weekend, I walked away with more than just a project. I met talented, driven people I am still connected with. I experienced the grind of building under pressure. And I realized I wanted to keep doing this: building intelligent, real-time systems that combine machine learning with thoughtful design. UofTHacks was not just my first hackathon. It was the moment I understood how much I enjoy pushing myself outside my comfort zone.