An AI agent for ISP operations
Can an ISP team get customer and network context without switching between dashboards?
An AI agent that retrieves customer, billing, and network information for an ISP’s operations team.
I like making things and figuring out how they work.
Full-stack engineer, engineering team lead, and SaaS founder. I build products, experiment with AI, and write about the stuff I learn along the way.

My parents had a plan for me.
It involved healthcare, vocational school, and a sensible chance of getting a job.
It did not involve me spending half my life in front of a computer.
Back in junior high, they wanted me to take the safer route. I understood why. From their point of view, it mostly looked like I was playing games. “Software engineer” wasn’t exactly an obvious career path.
Then I joined a national competition, spent a few very late nights with my laptop, and somehow won.
That helped my case.
A lot.
My parents started seeing the computer a little differently. They agreed to let me study Computer Science at university.
And apparently, that was the beginning of a much longer relationship with computers.
Since then, I’ve built software for schools, worked with product teams, and eventually started my own SaaS company. Years of actually building things gave me a solid engineering foundation — long before AI could write code for me.
Now AI is part of how I work, too.
It helps me move faster and build more. But figuring out the problem, questioning the solution, and taking responsibility for what ships?
That part is still on me.
More about meFrom writing code for others to building something of my own.
I’ve worked on software for schools, online shops, and personal finance. Now I’m building Eyzet, a SaaS platform that helps internet providers run their operations without the usual headaches.
Open the project notebookSome of that work takes the form of an experiment: internal AI tools, or a door that opens with face recognition.
Can an ISP team get customer and network context without switching between dashboards?
An AI agent that retrieves customer, billing, and network information for an ISP’s operations team.
How could a camera and a card work together to control a door?
A door-access prototype combining face recognition, a matching card, and Raspberry Pi hardware, alongside a separate mask-detection experiment.
Could I get my drone to follow my face using its camera and some Python?
An experiment connecting a drone camera, OpenCV face detection, and Python flight controls. I got the drone to follow my face.
This diary is where I’ll write down what I learn from the work and experiments.
My static site returned thousands of 404s for PHP files it never had. How I traced the requests to scanner VMs with Cloudflare analytics, and the WAF rules that stop them.
How a language model requests data from your application, how your code handles the request, and how the result becomes a useful answer.
Set up MikroTik web-proxy logging, send requests to a UDP receiver, and prepare the records for a searchable web dashboard.
There’s space here for other people’s ideas, too: books, articles, and papers, with my reading notes alongside them.
I’m making a home for books, articles, and papers, with space for the notes that stay with me. My first reading entry is still to come.
Open the reading journal ↗And here’s what I’m working on today. Updated Oct 9, 2026.
Running my own SaaS and leading an engineering team, with hands-on work across code, automation, and CI.
Learning notes will go here when they’re ready to share.
My first reading entry is still to come.
Collecting questions for the playground.
Have an idea, a question, or something you’re working on?
I’d like to hear about it.