ParkHere
Smart-camera parking detection that pivoted into industrial truck tracking.
A class project that wouldn't stop
ParkHere started in a course co-run between PUC Chile and UC Berkeley's Center for Entrepreneurship and Technology. Massive class, interdisciplinary teams, real budget on the line. Our team pitched a smart-camera system that watched parking spots and streamed their status to a mobile app. Frictionless parking instead of circling for ten minutes. We won $7,000 USD in seed funding to build it.
What we thought would be a one-semester project ended up running for two years, winning awards, putting us on Chilean TV, sending us to Boston, and eventually opening conversations with Walmart Chile.
The smart-camera unit
Each unit was a Raspberry Pi with a camera, running the image processing and machine learning that decided whether a spot was occupied. The Pi pushed status updates to a backend, which fanned out to the mobile app so drivers could see in real time which spots were free.
The camera also had to live outdoors, so we designed and 3D-printed a protective case for it, iterating the design so it could hold the Pi and keep the camera at the right angle.
The Bridge to Boston
After the Engie prize, we were selected for The Bridge Boston scholarship and spent a month embedded in Boston's startup ecosystem, with access to networks within the MIT community. We refined the prototype and the business model, and represented PUC's School of Engineering at events around the city, including a Chilean innovation event organized by ChileMass at Venture Café Boston with CORFO and the Chilean Ambassador to the United States.
The technical work was secondary. What we got out of Boston was a different understanding of what a hardware startup actually has to do beyond the prototype, something no Santiago classroom was teaching us at the time.
From street parking to truck yards
Back in Santiago, we started conversations with Walmart Chile about a different problem: tracking the trucks moving through their distribution centers, with theft prevention as the underlying concern.
We adapted the system into ParkHere Enterprise: the same vision core, pointed at trucks instead of parking spots. The pivot taught me that the technology was the easy part. The work was in listening carefully to a customer whose problem looked superficially similar to ours but had completely different constraints.
What it became
ParkHere was my first real lesson in building hardware for a real customer. It started as a class project, won prizes, and traveled further than we ever expected. It taught me how to take a research-grade idea and stress-test it against the world.
It's also where I started cultivating an edge-first instinct: keep processing as close to the sensor as possible, send only what's necessary upstream. That instinct came back over and over in the PhD work that followed, and it's still how I think about systems today.
