Pull up a chair, the coffee’s on.
I’m Andrei Migatchev — technologist, builder, and serial co-founder. This is my lounge: a place for notes on building AI that behaves, security that holds, and teams that ship. Stay a while, poke at things, send a message or invite me for coffee.
Curious types press `
The short version
I’ve spent my career in the trenches of technology — co-founding and building companies in healthcare AI, regulated fintech, and digital identity, and taking them from a whiteboard sketch to systems running in production across several continents. The common thread: I build things that most find too challenging or even impossible to create.
These days my work sits where AI capability meets accountability — designing and shipping LLM-based systems, agentic orchestration, and the evaluation and assurance machinery that lets regulated organisations trust them. Underneath that is an old habit from offensive security: assume it will be attacked, misused, or quietly drift — and design for that day first.
I’m the hands-on kind of technical leader. I still write code, still read the logs, and still believe the best strategy conversations happen close to the keyboard. If that sounds like your kind of conversation, the coffee pot’s never far.
How I think about shipping AI
I truly believe that every system built should live or die by this loop. Click a stage to drop it and see what actually happens. (It’s never pretty.)
The loop holds.
Build, secure, ship, monitor, learn, repeat. Boring on purpose — boring is what production is supposed to feel like.
click a node to toggle it ·
When there’s an interesting idea…
I’ve got to try it out straight away — and see whether it actually works.