I labelled the training data before I ever built with the models.
I'm Tori. I write about building with AI and learning how it actually works, from Nairobi. Here's how I got here.
From a sales floor to building with AI.
I live in Nairobi and I work online, but I didn't start there. I got here sideways, through a string of jobs I took because they paid, and one decision that changed everything.
Sales floors at twenty
I started in sales, working physical stalls, pushing the products nobody else wanted to push. You stand in front of strangers all day, you get told no, you keep going anyway. I walked out of it with a different relationship to my own confidence.
Whatever paid
Then I moved online. Transcription first, then essay writing, the way a lot of people in Nairobi get started on the internet. You take what pays, you watch accounts close without warning, you spend long stretches broke. What I kept from those years was tenacity, and one realisation: if I wanted to make it, I had to upskill on purpose.
Labelling the training data
In 2021 I spent a year labelling training data for large language models. At the time it was just work. It turned out to be the most useful year I have had, because I saw how these things are assembled and where the seams are, years before anyone was selling them to me.
Reddit told me to write
So I asked Reddit what was marketable, and copywriting kept coming up. I joined a Discord called the Copywriting Collective and I'm still grateful to those people. Office hours, copy annotated line by line, teardowns I sat all the way through. That community is where I actually learned to write.
An Australian agency
From there I landed a content writing role at an Australian agency, mostly content marketing, and stayed two years. Long-form, SEO-heavy work across a lot of different industries, with high standards. That is where the craft got real.
The full stack
After that, a US-based digital marketing agency, and the full stack of it. I rotated through more roles than I can cleanly list. I like to be challenged, and they kept letting me move: pulled into a department as the fresh pair of hands for some new process, get it working, write the SOP, get moved again.
Search, then systems
Today I do SEO and content operations for a portfolio of personal brands, and I build automation systems on my own time.
Going down a layer
I have built and handed over three multi-agent pipelines, and I can tell you what each one costs per run. That turned out not to be enough. So I am going down a layer, to the maths and the machinery, and writing down what I find on the way.
AI as a teammate
A lot of it runs through Claude Code: research, building the tools, rendering reports and presentations, putting the SEO newsletter together. The craft now includes directing the machine well, and knowing what is worth keeping.
"I do my best work when I'm put on something I haven't done before." What I've learned about myself
Organic isn't free. It's the most misunderstood phrase in SEO, and quality keeps widening the gap.
Content drawn from real work outranks content drawn from other content. Almost every time.
I'm not the loudest voice in this space. I'm in it every day, on real sites, watching what moves.
Flow Blueprints
How AI systems actually get built, run and handed over. I take one thing apart at a time: a system, a procedure, or something I got wrong. You get what it cost, where it leaked, and what I would hand to somebody else.