Computer vision
I train and evaluate detection models with YOLOv8: preparing datasets, reading precision and recall, and improving from there.
Used in SPAI and WilderCards
Software engineer. I build things that work in the real world, from vision models to the hardware you tap.
Computer vision, apps for Apple Vision Pro, mobile backends and NFC hardware. Every project starts with the person using it.
Not for riches. Not because it's possible. Because it helps.
Points first, the way Vision Pro sees a room. Then me.




A Vision Pro assistant that helps catch mistakes before surgery, using vision models I trained.

Snap a photo of any animal and get a trading card for it. I built the whole Firebase backend.

Tap an NFC card on an ESP32 reader, and attendance logs itself, end to end.
More on the way. Follow the builds on GitHub.
I train and evaluate detection models with YOLOv8: preparing datasets, reading precision and recall, and improving from there.
Used in SPAI and WilderCards
Apps for Apple Vision Pro and iPhone in Swift, SwiftUI and RealityKit, with Core ML for running models on the device.
Used in SPAI
Android in Java, React Native, and the services behind them: Firebase Auth and Firestore, REST APIs in Node and FastAPI.
Used in WilderCards and TouchPoint
Microcontrollers talking to real software: an ESP32 and PN532 NFC reader that turns a tap into a logged record.
Used in TouchPoint
The fundamentals under all of it: Python, C++, JavaScript and SQL, socket programming, database design, Git and clear documentation.
B.S. Computer Science, Full Sail University
The same four steps run through every project, from a vision model to an NFC reader.
Before any code, I figure out who is on the other side: a sterile processing tech, a student tapping in, a kid photographing a turtle.
I get a rough version working all the way through early, like TouchPoint's first tap landing in the database.
Precision, recall, a test that passes. I want numbers that tell me whether it really works.
I document as I go and share what I build in public on Juan Builds Tech.
Robotics, customer engagement, a computer science degree, and years of fast-paced service work.
Not for riches. Not because it's possible. Because it helps.
I want to build medical and computing technology that moves the whole industry forward for the good of people. If your team is working on that, let's talk.