ABOUT

Cloud & AI,built for production.

A systems-minded engineer who has tracked the modern AI arc since GPT-2 — and ships the kind of infrastructure that runs while you sleep.

By Martin Pernikov — Cloud & AI Developer, Pleven (BG) / Lyon (FR)·Last updated:

Who

Who I am

I'm a Cloud & AI Developer working at the intersection of multi-agent systems, self-hosted DevOps, and full-stack engineering. I've been actively tracking large language models since the release of GPT-2 in 2019 — six-plus years of perspective on where the models came from and where they're going.

My day-to-day combines hands-on infrastructure (Docker, Linux, reverse proxies, multi-cloud) with applied AI: production agent harnesses, workflow automation, API orchestration across multiple providers, and the unglamorous plumbing that makes any of it actually reliable.


How

How I work

I move on a tight loop — Analysis → Plan → Execution → Verification → Documentation — and I keep that loop closed. Nothing ships on "should work." Every problem ends in a written-down procedure so the next instance of it costs less than the first.

I trust but verify: I give autonomy where it's earned and check the critical surfaces (secrets, configs, blast radius) myself. I prefer the simple solution that survives the next outage over the clever one that doesn't.


Strengths

What I bring

Systems thinking

Architect ecosystems, not isolated tools — agents, data, infrastructure, and content as one coherent stack.

Technical depth

Senior-level DevOps, Linux, Docker and backend; comfortable end-to-end from kernel to UI.

Resource efficiency

Cost-to-value first — pick the cheapest model, provider or pattern that meets the bar; failover the rest.

Critical thinking

Skeptical by default. No guessing, no autopilot — show the evidence, then act.

Communication

Public AI speaker in BG, EN, and FR; able to translate deep technical work for non-technical audiences.

Documentation-first

Every fix becomes an SOP, every secret has a procedure, every system is reproducible.


Philosophy

Build to scale, not to demo.

Most "AI projects" are demos in disguise — they fail the moment the operator stops watching. I build systems that hold up unattended: secrets hardened, providers failed-over, logs that actually help, and procedures any future operator can follow without me in the room.


Public Speaker

Teaching applied AI

I deliver public lectures in Bulgaria on how modern AI is reshaping daily life and work — most recently for 8th-grade students at PDTG "Dimitar Hadzhivasilev" during EU Code Week, and at the Svishtov Youth Center certification event. Trilingual: Bulgarian (native), English, French.


Want to work together?

Open to collaboration on cloud, AI agents, automation, and infrastructure.

Back to home

© 2026 Martin Pernikov · Cloud & AI Developer · Pleven, BG / Lyon, FR