$ cat ./about.md

ABOUT

The long version: how game servers and Photoshop turned into leading engineering teams, without the mix ever going away.

born 1997
based Vitória da Conquista, BR
writing code ~10 years
leading teams ~5 years
largest team 30 people
first language Pawn
01 / ~2000 to 2015

Before it was a job

I was on a computer from about age three, and games are what pulled me in. The first code I wrote was for GTA San Andreas multiplayer servers, game modes in Pawn, around age twelve, with no idea what a variable was. I read other people's scripts, changed one thing, watched what broke, changed it back. That was the whole method for years.

Photoshop arrived at the same time, not after. CS1 and CS2, forum signatures with the lens flares and the bevels, then Dreamweaver layouts sliced into table cells. Responsive design did not exist yet.

Hosting those servers meant renting a US VPS, SSH with no graphical anything, and shell scripts for whatever I got tired of typing. Infrastructure before I knew the word. I still run a home server today, mostly to break things on.

02 / 2015 to 2019

School, and the detour

Computer Science in Salvador. I went in wanting games and spent the early semesters on OpenGL and toy engines, then dropped it because I could not see a path: in 2016, from Salvador, nobody I knew had a games job, and I was not confident enough to bet the decade on being the first. Front-end looked like something I could actually get paid for. It was.

My thesis went the other direction entirely, computer vision and machine learning with PyTorch and OpenCV. I never became a Python specialist, but it is still what I reach for when I need a script rather than a product.

The part of this stretch nobody expects: from 2017 to 2019 I studied aviation, mostly theory with some practice, and came close to switching careers to it. I did not. It faded out over a couple of years instead of ending on a decision.

03 / 2016 to 2021

Selling the work

Freelancing was the first money. WordPress sites mostly, some Shopify and WooCommerce, plus logos and social media design, because a client who needs a site usually needs six other things and I was not going to say no to any of them.

My longest client stayed with me for years. I was the only developer on their site and internal systems, and also the designer, the social media person, and for a while the photographer at events. If you want to learn how to scope work and say no later in your career, that is a good place to learn why.

Then 2020. I used the pandemic to properly learn modern front-end, React and everything around it. It is the only time in my life I have had months to study one thing without a deadline attached.

04 / 2021 to 2023

An agency, then my own

Vasta was a Shopify agency, where I learned Shopify for real and learned how an agency actually runs. An agency sells hours, so the job under the job is managing them: estimating, splitting them across projects that have nothing to do with each other, noticing when scope quietly eats them. I was on the build side: stores from scratch, visual work, the long tail of tuning, some Python automation and Selenium tests, and near the end UX and UI design for a Shopify app.

Then I co-founded EcomNinjas with a partner from Vasta and ran it for about a year. I was the one who drifted toward leading rather than building, which I did not notice at the time and only makes sense looking back.

05 / 2023 to now

Leading

I joined ilegra on client work, first Biossance, whose parent company is Amyris. Front-End Developer on paper, Technical Project Manager for a few stretches. Leadership was not my job there, but it is where I started studying it up close: I sat near my boss and the stakeholders, watched how decisions actually got made, and got to practice small pieces of it. After Biossance I project-managed Dexco, an infrastructure and DevOps team, and Gerdau US, large-scale data analysis with business rules that took weeks just to write down. Both went well, and I am still proud of the feedback from both.

Then I left and took a break. I helped people build e-commerce informally and finally had time for Three.js and WebGL. I came out of it convinced that 3D and creative interfaces are an opening in e-commerce, specifically for premium and niche products where a standard conversion-optimized layout does nothing for the perceived value of the thing you are selling. I keep betting a little on it.

Brainstorm came next, short but dense: my first formal tech lead title. I restructured an existing team, redesigned how Shopify themes were built, talked to clients directly, and reported to the CTO and the CEO. And I coded a lot there, more than people assume the title allows. Architecture, client calls, and pull requests in the same week, which is still my favorite shape for the job.

Now I am back at ilegra, overseeing three tracks that have nothing to do with each other: teams building AI products, teams building sites on Webflow, and e-commerce teams on Shopify. Since 2025 I have been studying agentic programming daily, safe AI coding practices and spec-driven development in particular. In practice that means LangChain, n8n, and Amazon Bedrock including Agent Core, applied in architecture and in prototypes I build myself.

How I lead
Principles and the general goal stay fixed. The specific way you get there has to change with the person and the situation.

I have led people with fifteen years of experience and people on their first week. The same approach does not work on both, and pretending otherwise is how you end up micromanaging one and abandoning the other.

The throughline

Jack of all trades is my own word for it. Game servers and Photoshop as a kid, logos and WordPress as a freelancer, architecture and people management now, still prototyping with my own hands on weekends. I never chose this as a strategy. It is closer to a temperament, and the main thing it costs me is depth in any single one of them.

Now
iuri@io:~$ ps aux | grep iuri