I'm ⚡️ Manuel ⚡️, an "artisan" in the programming domain. I see and create programs not only as a deadline to meet, but a canvas to paint for peer engineers to critique.

I obsess over simplicity, structure, and clean architecture. I optimize for readability over complexity, don't repeat myself in code, functions have a single responsibility, and I commit atomically.

While my recent work anchors on AI-driven software, I craft with an eye toward edge-level performance and automated, friction-free DevOps pipelines.

• EyeZite - Our undergraduate thesis. A diabetic retinopathy (DR) lesion detector and grade classifier, with me as the project lead and backend developer.
• vance - A unified CPU scheduling simulation Python package, published in PyPI. Got tired of scattering implementations among the internet. Included hardware representations as well.
• Debunkr - A Filipino fake news classifier, with ensemble learning. I led the project, engineered the backend, manually crafted the preprocessing pipeline, and trained the AI models.
• bud - A Git-style, lightweight, keyboard-driven personal budget tracker for terminal users
• Rezbin AI - My internship project from Rezbin, a multiclass trash classifier using MobileNetV2, trained with selective layer freezing. Deployed as a microservice using FastAPI with Docker. Primarily a paper replication

• WVSU Re:Claim - A lost & found web app for WVSU, founded & led by me. Managed dev, marketing, operations, and creative teams.
• SPARK Hub Website - A fullstack website for my most beloved innovation student organization. Led the entire project, but focused on CMS development.
• BINHI Website - WVSU's official technology business incubator website, with me as one of the frontend developers.

• ✉️: carado.johnp11q@gmail.com
• 🔷 LinkedIn: John Manuel Carado
• ✒️ Blog: https://jmanns.bearblog.dev