About
I build software and show how
I am Rodrigo. HustleTech is my company, and right now HustleTech is also just me. Fifteen years of fundamentals and real fluency with AI: not the vibe coder who ships without a base, and not the veteran who thinks AI is autocomplete.
The record
Numbers that can be checked, not adjectives.
15+
Years in the industry
2
Platforms in build
16
Repositories
3
Languages published
One person
There is no team here, and that is the point.
You will not find a "we" anywhere on this site, because there is nobody else. Sixteen repositories, two platforms, three languages and a portal that costs twelve dollars a month, kept standing by one person who also holds down two jobs.
If that sounds impossible, it is because most of what gets attributed to team size is actually coordination. Tools do the boring eighty percent, AI included, and I say exactly where on the methodology page. The judgement is mine, and the judgement is the part I sign.
The position
Two people are wrong about AI, and they are wrong in opposite directions.
The vibe coder
Ships fast, does not know why it works, and has nothing to stand on the night it breaks in production. The output looks the same right up until it does not.
The veteran
Knows every fundamental, decided AI was autocomplete for LinkedIn, and is being passed by people who understand less than he does.
Where the model fails is exactly where the fundamentals matter.
That is the whole argument, and it is why I do not assert any of it. A model writes hreflang that is syntactically perfect and points every canonical at the English page, which silently kills the entire cluster. Me catching that is not a claim about competence. It is competence, on screen, in a video you can go and watch.
The road
Fifteen years, and the part that compounds.
2011
First line shipped for money
Start of the industry clock. Systems analysis and infrastructure first, and everything after this compounds.
2016
Founded HustleTech
My own software company, architected from zero, running in parallel to everything else.
2019
Moved it to São Paulo
The company follows the work, and the work moved.
2024
Banking and fintech
Java, Spring and Python at a Tier-1 payments provider, on systems where people and money quietly depend on the thing staying up.
2025
AI-native, in production
LLM-assisted incident triage built on my own initiative and taught to the team. Average triage went from about an hour to about twenty minutes.
2026
Building in public
Local-first GenAI on my own hardware, and this portal, with every decision on camera.
Method
Solved first, filmed second, and nothing invented in between.
- Video is the source. The article is derived from it, translated, and reviewed by a human in every language before it publishes.
- AI is in the loop and it never gets a byline. Authorship carries accountability and a tool cannot hold any, so every piece is signed by me and the tooling is declared once, on the methodology page.
- The part that broke stays in. Cut the difficulties and sell ease and it stops being useful to anyone.
- Errors get corrected in public, dated, and propagated to all three languages.
The work
Where being wrong is expensive.
Banking and fintech
Brazil, US, Europe
Java, Spring and Python on systems people and money quietly depend on. Ten years of the kind of bug that costs money when you miss it.
AI-native, in production
Tier-1 payments provider
LLM-assisted incident triage: about an hour down to about twenty minutes. It investigates and recommends, a human decides and acts.
This portal
Built in public
Static site over a database, three languages, $12 a month. Every decision is a video.
HustleTech News
Every Thursday, what did not fit in the video.
One edition a week on money, technology, work and discipline. No feed, no noise.
No spam. One-click unsubscribe.