Skip to content
peresramirez

❯ peresramirez.com

Don't let AI leave you behind.

I'm Javier Peres, an Electrical and Software Engineer. I teach you how companies are actually using AI inside their processes: where, what for, and with which tools. Because there is so much content, all of it equally promising, that it has become hard to tell what actually works.

About me

A real terminal, not decoration.

Type a command or press one of the buttons. Start with “whoami” to find out who I am, or “why” to find out why I do this.

peresramirez — zsh
❯

What I do

01

Using AI in your work

I teach you to use the tools that already exist for the job you already do: writing, analysing, summarising, deciding faster. Without building anything from scratch, and without needing to code.

02

Integrating it into your company

Connecting the apps you already use, automating the steps that repeat, and leaving workflows running on their own. Starting from the process that hurts, not from this week's tool.

03

What isn't worth it

I also say when AI adds nothing. A channel that recommends everything recommends nothing, and knowing what to discard saves more time and money than any new tool.

Why I do this

There's no shortage of information. There's a shortage of judgement.

I've seen up close how companies work in developed countries. They aren't trying AI out to see what happens — they have it inside their processes, and it earns them a real, measurable advantage every day.

The problem isn't a lack of information: there's too much of it. So much content, so badly organised and so full of promises, that telling what works from what merely sounds good has become a job in itself. Meanwhile the advantage goes to whoever is already applying the parts that hold up.

That's what I do here: filter. Teach what I've watched work, say plainly what isn't worth it, and learn alongside everyone else, because I'm still learning too. AI is the first wave. More are coming, and robotics is next.

live

Feed

Latest from me.

Videos, project stories and notes, newest first.

Thousands of small simulated robot ducks practising at once on rows of identical obstacle coursesWatch
Video2026-10-26

Microduck: I rebuilt NVIDIA's robot-training workflow with one AI agent

At GTC Berlin 2026, NVIDIA showed AI agents planning and preparing everything a small robot needs to learn to walk an obstacle course. I rebuilt that workflow with one agent, Claude Code, and a GPU rented by the hour. In the final test, the share of robot ducks that finish the course went from 39 % to 98 %.