Software · interfaces · systems

On charlles.dev, Next.js talks to GitHub, video responds to scroll and every project exposes the problem, decision and code. I’m Charlles Augusto, a full-stack software engineer in Campina Grande, Brazil.

Charlles AugustoFull-stack software engineer
Scroll to explore

Public work, real context

Open code. Visible decisions.

No project appears here as a stack alone. Each one names the problem, technical decision, next step and the repository where the work can be checked.

  1. WebTypeScript

    charlles-dev-portfolio

    A Next.js portfolio with public GitHub data, three languages and video controlled by scroll.

    Problem
    Show full-stack work without reducing every project to a polished image and a technology list.
    Technical decision
    Fetch public GitHub data on the server, localize the narrative in three languages and synchronize video, content and scroll without blocking reading.
    View project
  2. InfrastructureGo

    Astrolink

    Infrastructure and software to sell and manage Wi-Fi access on local networks with Starlink.

    Problem
    How to bring a low-cost connectivity operation to remote areas and communities.
    Technical decision
    Separate a Go backend, a SvelteKit captive portal and integrations for payments, vouchers and OpenWrt.
    View project
  3. ExperimentsKotlin

    trakr

    A smart tool case that identifies equipment through RFID/NFC and connects an ESP32 to a Kotlin interface.

    Problem
    Detect a missing tool before the team closes the case and leaves the worksite.
    Technical decision
    Distribute identifier reading to the ESP32 and centralize inventory, states and alerts in the Kotlin application.
    View project
Charlles following a broken cable through the dark with a flashlight

About my work

I start with what must work. The visual layer comes with it.

I move between interfaces, backend and automation. Here that means GitHub data, three languages, accessible states and media that never blocks the reading. Visual detail does not hide the code; it explains what the system is doing.

  • The interface leaves clues

    Hierarchy, states and motion show what happened and where to go next.

  • The ends must talk

    APIs, data and automation belong where they remove a manual handoff or a user doubt.

  • Failure is part of the build

    Loading, errors, keyboard use, slow connections and reduced motion are handled before release.

View my LinkedIn profile

Professional practice

Experience that leaves a trail.

A timeline of verifiable projects and responsibilities. Formal roles, clients and metrics only appear when they can be published accurately.

  1. 2026 — present

    Full-stack engineering and direction for Charlles.dev

    Architecture, interfaces, public integrations, localization, SEO and motion.

    A public portfolio with verifiable code, decisions and a quality gate.
  2. 2026 — present

    Initial architecture and software for Astrolink

    An investigation into low-cost connectivity for remote areas.

    A public proposal and Go base that keeps pending validation visible.
  3. 2026 — present

    Hardware and app prototyping for Trakr

    Physical inventory using RFID/NFC, ESP32 and Kotlin.

    A product hypothesis turned into a traceable integrated experiment.

Outside GitHub

The code shows what I built. This is what I studied to get there.

Cybersecurity with Hackers do Bem, AI foundations through IBM SkillsBuild and computer vision at Geração Caldeira. Certificates and tools provide context; they do not replace practice.

Education

Study, test, break, understand. The order varies.

I use these learning paths to close gaps that surface in projects: networks and risk, AI foundations and image analysis.

  • Security

    Hackers do Bem

    Training focused on cybersecurity fundamentals and a more critical reading of systems, networks and risk.

    Cybersecurity / Networks / Foundations

  • Artificial intelligence

    IBM SkillsBuild

    Study of AI fundamentals, its uses and the care required to apply the technology with context.

    AI / Data / Application

  • Computer vision

    Geração Caldeira

    A practical path through images, models and ways to turn computer perception into a solution.

    Computer Vision / Python / Models

Certifications

Certificates help. Knowing where to apply them helps more.

They record parts of the journey. Their value appears when the content improves a technical decision, a review or the next question.

  • Geração Caldeira

    Computer Vision

    Foundations and practice for image-analysis solutions.

    Computer Vision / AI

  • IBM SkillsBuild

    AI Fundamentals

    Core concepts, applications and implications of artificial intelligence.

    AI / Foundations

  • Cisco

    Endpoint Security

    Device protection, threat vectors and defense principles.

    Endpoint / Security

  • RNP / Softex

    Hackers do Bem

    Introductory training and foundation building in cybersecurity.

    Cybersecurity / Networks

Stack with proof

A tool earns its place by solving something.

Select a technology to see where it appears and what responsibility it carried.

Localized routes, rendering, metadata and server/client composition for the portfolio.

Evidence in

Let’s talk

Where is the system getting stuck?

Send the context on WhatsApp or book 15 or 30 minutes on the calendar. No form and no canned pitch.

What should we talk about?
Direct messageChat on WhatsAppIf you know what must be built, or can only point to where it hurts, send the context. I read it before replying.15 or 30 min · online callSchedule a callSend the context on WhatsApp or book 15 or 30 minutes on the calendar. No form and no canned pitch.
Open résuméhello@charlles.devDiscord
© 2026 Charlles AugustoBuilt with Next.js and attention to detail.