About
The story behind Nubes Academy.
About Nubes Academy
The gap between knowing technology and operating it at production scale – I’ve spent 20 years watching engineers struggle with it. That’s why I built this.
Nubes Academy is a structured technical education platform focused on Machine Learning Operations – the discipline that determines whether an organisation’s AI investments ever reach production, or stay in notebooks forever.
I’m Alex Pototskyy, the founder. I’m an independent IT consultant with over 20 years of experience, and I built this because the education I needed for the engineers I was advising didn’t exist in the form they needed it.
Alex Pototskyy — Founder & Senior IT Advisor
Where it started
In 2020, I was teaching DevOps and Cloud topics to Master’s students at a university in Ukraine. It was the kind of teaching that reminded me why I got into this field: brilliant people, serious about their careers, trying to understand systems that are genuinely complex.
But every cohort showed me the same problem. Students understood the theory well — they could describe CI/CD pipelines, explain containerisation, discuss monitoring architectures. What they couldn’t do was build anything that worked in production. The gap between academic knowledge and operational reality was enormous, and the existing courses weren’t closing it.
“The problem wasn’t ability. It was that nobody was teaching production engineering the way it’s actually done — starting from a real system, learning the minimum necessary to make it work, and building from there.”
Nubes Academy started as an attempt to solve that specific problem. Not a broad catalogue of technology courses, not a certification factory, not a platform that teaches tools for their own sake. A structured, practical path that trains engineers to operate real systems — and does it in the sequence and depth that production engineering actually requires.
Five years later, 29,000 engineers across 190 countries have taken that path. The gap I saw in that university classroom in 2020 is still real. The MLOps Specialization — eight courses covering everything from the model lifecycle to security and compliance — is my most complete answer to it yet.
The instructor
20 years of production engineering. Every course reflects it.
My career started in system administration, moved through software development, and eventually into architecture and technical leadership roles across companies and countries. Along the way I worked with teams that were shipping models to production — and watched most of them make the same preventable mistakes in the same predictable sequence.
I’ve been an independent IT consultant for years, advising startups and enterprises on the technical decisions that determine whether their ML systems work reliably or fail silently. That advisory work runs parallel to Nubes Academy — it’s where I see, in real time, what skills organisations need and can’t find, and what the engineers they’re hiring are missing.
Teaching at university level gave me the second important input: knowing how to structure knowledge so it transfers, not just how to demonstrate it. The combination — active practitioner who also teaches at curriculum depth — shapes every course I build.
How we teach
Four principles that separate production engineering education from everything else
These aren’t marketing claims — they’re specific pedagogical decisions that every course is built around. Each one comes from watching what actually works when you’re trying to train engineers who will own production systems.
Minimum necessary knowledge — not exhaustive coverage
Every course is scoped to the minimum knowledge required to be effective at the next level. Topics that are important but belong in a different course are explicitly deferred with a note explaining where they’ll appear. This prevents the cognitive overload that makes most comprehensive courses ineffective in practice.
A single running project — not disconnected examples
The MLOps Specialization uses the same Customer Churn Prediction Service across multiple courses. You watch the same codebase evolve as each new concept is applied to it. This is how engineers actually learn production systems — by seeing how each layer connects to the others, not by studying them in isolation.
Market-validated curriculum — not invented from a textbook
Before the CI/CD for MLOps curriculum was finalised, every module was validated against 205 real MLOps job descriptions. Topics that appear frequently in senior job postings are taught in depth. Topics with no market signal are trimmed or removed. The course teaches what employers actually hire for — not what looks thorough on a syllabus.
Principles first, tools second — skills that don’t expire
Tools change. The MLOps Specialization teaches the operational principles that the tools implement — so when MLflow is replaced by something else in three years, the engineers who understand model registries conceptually can adapt immediately. Tool-specific knowledge is provided where it’s necessary; conceptual mastery is the goal.
The journey
Five years from a classroom idea to an international platform
Nubes Academy has been operating since 2020. This is what the growth actually looked like — not a summary, but the real sequence of decisions and milestones that got us here.
The platform in numbers
The 29,370 student figure reflects total enrolments across all courses and platforms. The 190 countries figure reflects the geographic distribution of that student base. The 205 job descriptions figure reflects the market validation methodology used for the MLOps Specialization curriculum design.
What’s next
The 2025–2027 build
These are the milestones being worked toward — some already underway, some planned. Honesty about what’s live versus what’s in development is part of how Nubes Academy operates.
MLOps Fundamentals, Python for MLOps, and ML Algorithms for MLOps are available now. Courses #4 through #8 are in production and releasing step by step through 2026.
Team licences for 5–20 engineers, with manager progress dashboard, kickoff session, and Q&A access. Learn more →
ML Model Lifecycle, CI/CD for MLOps, Data for MLOps, MLOps Observability & Monitoring, and Security & Compliance for MLOps. Each releases as it passes quality review. View the full path →
The Specialization Certificate is designed to grow into a recognized professional qualification through formal university accreditation. Partnerships are in early development.
A structured mentoring layer and an Income Share Agreement programme for qualified students who need deferred payment. Both planned for when the full specialization is available and outcome data supports the model. Register ISA interest →
A contextual AI assistant that answers student questions in the context of the specific course module they’re studying — giving every student access to the kind of on-demand guidance that currently only exists in expensive one-on-one tutoring.
Stay connected
New courses.
Better price.
Subscribers first.
- Course launch notifications — when Courses #4 through #8 of the MLOps Specialization go live, subscribers hear first and receive subscriber pricing
- MLOps field insights — what’s changing in the job market, what employers are hiring for, and what skills are becoming table stakes at senior level
- Free content and guides — practical material for engineers working with production ML systems, published when it’s ready, not on a content calendar
Stay in the loop No noise. No spam. Unsubscribe any time.
Your data is stored securely and never shared. · Privacy Policy