01
Oil Palm Fruit Count System
Real-time palm fruit tracking using YOLOv8, an ESP32 display, and a Django interface. A practical bridge between AI and agriculture.
Where thoughts take shape and quietly shine.
Most of my work happens when the house gets quiet. Just the low hum of the fridge, a cup of coffee I forgot to finish, and the glare of a monitor.
I didn't start programming because I wanted to build grand things. Honestly, I just liked taking stuff apart to see what made it tick. But somewhere along the line, that quiet curiosity turned into code, and the code started talking to hardware.
Software on its own can feel pretty isolated. It's usually just abstract logic sitting on a server somewhere. I think that's why I'm drawn to AI and hardware systems. I like building small bridges between the digital and the physical. A camera trying to count palm fruits out in the dirt and heat. A sensor noticing the air quality in a room.
It's rarely a clean process. There's a lot of messy trial and error. You spend days fighting a model that refuses to learn, or staring at a breadboard trying to find a loose wire. You start to wonder why you bother.
But then the script finally runs. The data comes through. The bridge holds. For a brief second, it feels like the machine actually understands the room it's sitting in.
It's a very quiet kind of satisfaction. I'm Julio. I write code, I wire sensors together, and I try to make a little bit of sense of the noise.
Squeezing vision models onto tiny devices, bringing a little bit of sentience to overlooked corners.
Untangling messy workflows and building quiet connections between isolated systems.
Piecing together enterprise modules and navigating the heavy, structured world of ERP architecture.
Real-time palm fruit tracking using YOLOv8, an ESP32 display, and a Django interface. A practical bridge between AI and agriculture.
A smart waste-sorting bin powered by YOLOv8 and an RFID reward system. Making recycling feel a little more like a game.
Automated potato counting on a live webcam feed. Built with YOLO, a Tkinter GUI, and automated database exports.
A CV evaluation app powered by ML and NLP models. Built as a capstone project during Bangkit Academy.
A quiet, real-time environmental monitor reading air and waste metrics via ESP32, visualized on a Django dashboard.
A Java Swing ledger for costume rentals. Handles inventory, patrons, and late fees under a strict MVC architecture.
There's a strange, quiet thrill in watching a model try to predict reality out in the wild. It's a bit like watching a child point at the moon for the first time: flawed, maybe, but deeply meaningful.
#computer-visionJumping between hardware and software taught me that the best systems are just bridges. Every layer you build is just another translation, turning raw data into something human.
#iotMost of the things I build will never see the light of day. They're just abandoned experiments and blurry 3 AM ideas. But the strength of an iceberg always lies in what you can't see.
#processA flocking simulation built around Craig Reynolds' Boids algorithm. Every particle only knows three simple rules: separation, alignment, and cohesion. Put them together, and they start to feel alive. Move your mouse to give them a little direction.