I started messing with software development in 2013, purely as a hobby, and decided fairly early that it would be my profession. I was 10 at the time — that's 12 years of writing code.
I've worked at several companies since, mostly in web development, though the job has also taken me through mobile apps and desktop solutions. Along the way I've met technologies from this decade and technologies older than I am, and both taught me something worth keeping.
What I enjoy most is the part underneath the framework: why a data structure behaves the way it does, what a protocol actually asks of you, where an abstraction stops paying for itself.
This blog is where I write that down, in English and Portuguese, mostly so I'm forced to understand it well enough to explain it.
The path from a Minecraft server to writing software for a living.
At 10 years old I wanted custom systems and commands on my server, so I learned logic, events, variables and conditions inside Skript. That was the gateway — and where I found out I actually enjoyed this.
SkriptLang/SkriptWanting to go past what Skript could express pushed me into writing actual plugins, and from there into the language itself: types, the JVM, and my first real contact with software design.
The hobby became the profession. Web development took the front seat, but the work also stretched into mobile apps and desktop solutions.
Clipper was released in 1985 and has no abstraction for any of this, so building a complete API from scratch meant reading the RFCs and understanding, byte by byte, what actually happens when a request reaches a server.
Software that aggregates sales from thousands of markets and puts the numbers on the client's phone in near real time. The hard part was working around the limits of the legacy stack by making it coexist with newer technology instead of rewriting everything.
An orchestrator for Robotic Process Automation with a log of everything that happens, from the trigger to the end of the run. Managing, starting, stopping and auditing each automation stopped being guesswork.
Software that scores how safe it is to extend credit to a given customer, using AI to read their documents, registration data and history before it answers.
A feature that streamlines reporting defaulters to the Brazilian credit protection bureaus, and clears the record automatically the moment the customer settles the debt.
Software that runs collections on outstanding invoices: who has been contacted and through which channel, status updates when payment lands, and generation of the payment methods most used in Brazil.
Started as a place to write down what I was studying in depth — data structures, Go, the things that are easier to understand once you have to explain them. Bilingual from day one.
I don't marry a tool — the domain is what carries over. Here's what I work across, with the tools I've used as examples, not a ranking.
TypeScript·JavaScript·Java·Go
Node.js·Spring Boot
React·Next.js·Tailwind CSS
Relational·Non-relational·In-memory / cache
OpenAI·Anthropic·MCP·RAG·Agents·Embeddings
Linux·Kubernetes·Docker·Vercel
Kafka·RabbitMQ
Grafana·Prometheus·Datadog·Sentry·OpenTelemetry
Git·Bash·GitHub Actions·GitLab CI
A few things that explain more about how I think than a list of technologies would.
I got into programming to bend a Minecraft server to my will. I wanted custom systems, commands and mechanics, so I picked up Skript and learned events, variables, conditions and control flow without ever realising I was learning to program. The feedback loop was instant and the motivation was entirely mine — a better teacher than any course, and the reason I later went deep into Java for real.
The Skript pluginClipper predates almost everything we take for granted, and it has no abstraction for HTTP — no framework, no router, no request object. Building a full RESTful API on top of it meant reading the RFCs and implementing the protocol by hand, byte by byte. It taught me more about how the web actually works than any modern stack ever could: when nothing is hidden, you don't get to skip the part where you understand it.
I'm genuinely fond of every tool I use — and I think that's exactly why I'm quick to point out their flaws. The more you know a technology, the more you see where it leaks. Not having a darling doesn't mean drifting, though: I made a deliberate bet on one stack — the TypeScript web ecosystem — to build a career on and go deep, while keeping the freedom to reach for whatever a problem actually needs. Right trade-off over trendy, always.
The most recent articles, if you'd rather read than scroll.
Questions, corrections or just to say hello — any of these work.
