AI SYSTEMS BUILDER · LEARNING IN PUBLIC
Automation, agents, and weird internet experiments built one system at a time and documented while they break.
// SELECTED BUILDS
Small systems with one rule: they need to solve something real, not just look impressive in a workflow screenshot.
AI AGENT × DECENTRALIZED IDENTITY
An AI generates a message in n8n, a local bridge signs it with a DID, Technocore records it, and the sequence proof returns to the workflow.
PROSPECTING ENGINE
Scrapes a business, understands its offer, detects lead-gen signals, scores ICP fit, and surfaces decision-makers.
LOOSE-ENDS ASSISTANT
Telegram capture becomes structured tasks, research, content ideas, build notes, and follow-ups with dates and next actions.
// DAILY STACK
I use whatever gets the idea from “this should exist” to a working test fastest.
// ABOUT
I’m learning what AI can actually do by building with it. Automations, agents, websites, trading experiments, and whatever else seems worth testing.
I write down what broke, what fixed it, and what the system taught me.
@zinitoh on X ↗// CURRENT STATUS
