Who’s behind it
It didn’t start today. It started 30 years ago.
ATAILA is founded and run by Attila Macskásy — a Hungarian infrastructure engineer and cloud architect who has built, sold and operated every layer this platform bundles. In this field credibility is earned in decades, not in prompts. Here is where his comes from — and each era, in hindsight, was training for this one.
-
1992 · The beginning
A Commodore Plus/4 and no games
As a kid he wanted to repair televisions — the family set kept breaking, and the repairman seemed like a magician. Then a cousin brought a Commodore Plus/4 from Vienna. It had almost no games, so he learned BASIC instead. That trade — understand the hardware, then program it — never ended.
-
1998 · 1998–2003
Ericsson — datacenters and web applications
He ran the server room for a thousand users, then built Ericsson Hungary’s intranet — a dynamic, database-backed web application, in 1998. Datacenter foundations plus full-stack software in one job: the combination stuck for life.
-
2001 · ~10 years
His own service provider
For a decade he ran his own hosting business — ASP, SQL Server, LAMP, VMware, Zimbra — while writing telco fraud-detection PL/SQL against one of the country’s largest Oracle databases. He knows the failure modes of operating infrastructure first-hand, not from a slide deck.
-
2009 · 2009–2013
VMware — the first VDI in Hungary
He became the first VMware systems engineer in Hungary — hired partly because he ran the country’s most-read virtualization blog. He brought Virtual Desktop to the Hungarian market, keynoted to 600+ people, and grew the Budapest VMUG into the largest in Central and Eastern Europe.
-
2013 · 2013–2014
Virtual Instruments — San Jose
A Silicon Valley startup, selling infrastructure performance management into the region’s largest datacenters. Dual SAN fabrics, Fibre Channel zoning, microsecond latency at enterprise scale — and a crash course in how venture-backed companies actually operate.
-
2014 · 2014–2021
Microsoft — cloud on hardware you own
On a worldwide hosting team he helped Central Europe’s biggest service providers put Azure technology inside their own datacenters — Windows Azure Pack, then Azure Stack, taking partners to Redmond for the launch. “Cloud on hardware you own” has been his theme for over a decade.
-
2021 · 2021–2024
The multi-cloud years
Independent first, then multi-cloud architect roles across Azure, AWS, GCP, OCI and VMware — the certifications, the statements of work, delivery to enterprise standards. And a blog, cloudmigration.blog, to share the research.
-
2025 · Now
Back to the metal — private AI
By day he helps keep a very large production platform running — distributed systems, DevSecOps, AI in production. Nights and weekends he builds AI supercomputers on his own hardware: GPU fleets, DGX-class machines, 200-gigabit interconnects, open models served on silicon he owns — measured, documented, failures included.
“Running AI offline provides both privacy and independence. Even if you pay for cloud AI services, you’re never entirely sure how much of your prompts and data are retained. With my setup, nothing leaves my lab.”
Written in September 2025, when the first machine was built. A year later the one desktop had grown into a fleet — and the thesis had only hardened. ATAILA is that thesis, operated for customers.
ATAILA is this career, productized
The company is the same shape as the résumé. Each layer of the platform exists because he has already run it — for an employer, for his own customers, or on his own hardware.
A decade running his own hosting company
Everything ATAILA operates comes with a written SLA — from someone who has carried the pager for his own customers.
What “operated” includes →Seven years putting Azure into providers’ own datacenters
ATAILA Cloud: the building blocks of a hyperscaler, on European hardware, under EU jurisdiction.
The sovereign cloud →Bringing Virtual Desktop to the Hungarian market in 2009
ATAILA Studio: a managed development workstation with a private AI coding fleet — plug in and start.
How Studio works →Intranets in 1998, full-stack ever since
The Factory: from an application’s description to running DEV, UAT and PROD environments with governed releases.
How the Factory works →AI supercomputers built and measured at home
Private AI on EU GPUs — priced and sized by someone who knows what a GPU actually costs per day, because he pays that bill.
Private AI for your team →The long version, with invoices
The founder documents the hands-on work in public — hardware builds with real costs, measured tokens per second and watts at idle, and the failures included, because that is where the learning lives.
privateai.blog — the full story and the AI supercomputer build logs →
cloudmigration.blog — a decade of cloud, and now the deliberate way out →
Put the 30 years to work
If you have a workflow that needs private AI and someone who has actually operated every layer under it — tell us about it. You talk to the person on this page, not to a sales team.
Tell us about your project →