Article content
As AI rapidly reshapes software development and digital skills become a global priority, //kood is taking Estonia’s alternative approach to tech education far beyond its borders. In this interview, CEO Lauri Haav reflects on scaling the model internationally and why adaptability, peer learning and curiosity matter more than traditional credentials in the AI era.
//kood started as a single school five years ago. What has driven its growth to nearly 10 locations today, and what have been the key lessons along the way?
When we opened in Jõhvi five years ago, we were testing a hypothesis: that you could teach world-class software development without traditional teachers, lectures or textbooks, and that this would work in a small Estonian town as well as in any tech capital. What we have learned since is that the demand is definitely there and growing, and it comes from two very different groups.
One group is young people who don’t fit conventional education, or those who are very certain of their path after high school. The other, just as important and actually more than half of our learners, is career changers — people in their late twenties, thirties and forties bringing engineering, finance, healthcare, logistics, or any other background you can think of, into software. These career changers are gold for employers. They arrive with deep domain expertise and life experience that a 20-year-old cannot have. Once they add programming skills, they become exactly the kind of cross-disciplinary engineers companies struggle to find anywhere.
The growth to nearly ten locations hasn’t been driven by us pushing outward; it has been driven by communities, municipalities and partners pulling us in. The biggest lesson is that you can’t franchise a school the way you franchise a restaurant. You need committed local people who deeply believe in the model and are willing to defend it when things get hard, because in education, things always get hard.
You’re now expanding to places like Odesa and Moldova. What makes your model exportable, and what have you had to adapt in new markets?
What makes the model exportable is that we are not transferring an idea, but transferring a full tech education platform. It comes in four parts: a continuously updated curriculum that reflects what employers are actually hiring for, the online study platform students learn on, a school operator playbook that codifies how to run the school day-to-day, and a setup model for building the local expert and employer community that surrounds it. With those four in place, we can typically take a new country from commitment to first-cohort launch in roughly 6 months.
In Odesa, we are opening in a country at war, which changes everything from physical infrastructure to how learners plan their lives. In Moldova, we are working with a country actively rebuilding its tech sector and aligning with the EU. What we adapt is the wrapper: language and local customs, partnership structures, employer connections, and how we integrate with the local education system. What we don’t adapt is the core: the conviction that students learn best by working together in peer groups, that talent is everywhere and that motivation matters more than prior credentials.
More broadly, what does this say about the potential of exporting Estonia’s approach to digital education? Where does it work, and where are the limits?
What //kood is doing is a slice of a much bigger Estonian story. For thirty years, Estonia has treated digital capability as core infrastructure, not just for government services or e-Residency, but for education. We put computers and the internet into every school in the late 1990s with Tiger Leap and built a public sector that runs on digital identity. Today, Estonia consistently ranks among the top in Europe in the OECD’s PISA assessments in math, reading, and digital skills. By the time someone joins //kood at 19 or at 42, they have already grown up inside that ecosystem. That cultural readiness is genuinely hard to replicate.
What can be exported is the playbook for getting there. Estonia has a quietly impressive track record of helping other countries digitise public services, identity systems, and e-government infrastructure, and tech education fits naturally into that toolkit. It works best where two conditions meet: a real economic need for digital talent and political willingness to break with traditional educational hierarchies. Ukraine and Moldova clearly fit that profile, as do many countries in the Western Balkans, Central Asia, and parts of Africa. Where it hits limits is institutional flexibility. If a country cannot accept that a 19-year-old without a diploma, or a 42-year-old former nurse, can be a competent software engineer, the model won’t take root, no matter how good the technology is. Estonia’s real export advantage isn’t that we are better at programming. It is that we are more comfortable letting outcomes outrank credentials.
AI is rapidly changing software development. How is it already reshaping what and how you teach at //kood?
AI is the most disruptive force to hit our field since the internet, and we have leaned into it rather than away from it. Students at //kood use AI tools every day. We don’t pretend they won’t have access to them at work, so we don’t ban them in school.
The misconception we work hardest against is that AI removes the need to learn programming. The opposite is true. The best way to think about AI is as a multiplier. If your underlying knowledge is strong, AI multiplies it — you become two, five, ten times more productive. But anything multiplied by zero is still zero. A student who never learned how programs work, how to read code, or how to design a system gets nothing useful out of an AI assistant. They generate plausible-looking nonsense and have no way to know.
So the curriculum has actually become more demanding, not less. We now develop learners across four interlocking areas: technical fundamentals — architecture, problem decomposition, system design, code review; meta-skills — learning how to learn, working in teams, navigating ambiguity; product and business thinking — understanding why software gets built and what makes it valuable to a customer; and AI literacy — directing these tools rather than being directed by them. Together, that combination is what separates a good engineer from a useless one in the AI era.
Given how fast things are moving, how do you design a curriculum today that prepares people for jobs that may look very different in two years?
You stop designing for specific jobs and start designing for adaptability. We don’t promise learners they will graduate as Java developers or DevOps engineers; those titles may not exist in the same form by the time they finish. We promise they will graduate as AI-ready product engineers who can pick up whatever the industry needs next.
That requires constant market research. We run regular conversations with our employer partners across Estonia and abroad: what they are hiring for now, what they expect to need in 12 to 24 months and what skills the candidates they are rejecting are missing. We track hiring trends, technology shifts and which graduates get placed quickly versus slowly. That feedback loop is the engine that keeps the curriculum current. Projects rotate based on what we see. Tracks get added — AI, cybersecurity, modern infrastructure — and others quietly retire.
Because students learn from each other rather than from a fixed teacher, knowledge within the school updates faster than any textbook could, the deeper answer is that we anchor the curriculum on the things that compound over a career rather than expire: meta-skills, product and business thinking, and AI literacy. Specific technologies will keep shifting. Those underlying capabilities don’t. And those are exactly what AI cannot replace.
On top of the multi-dimensional market research and peer-learning model, we also have a board of advisors, which we call the Kood/Fellowship. This body consists of tech leaders with deep knowledge and frontier operator experience who have pledged their time and expertise to provide continuous feedback on the path we are taking and keep us on the right track. At the time of writing, the fellowship consists of Priit Kaasik, co-founder and former CTO of Katana; Oliver Leisalu, technical co-founder of Bolt; Margus Niitsoo, technical founder of SALK; and Mart Roosimägi, Swedbank’s Head of AI and scaling.
Looking ahead, what skills or mindsets will matter most for developers in an AI-driven world, and how should education systems respond?
The skills that will matter most are the ones AI amplifies rather than replaces: clear thinking, system design, asking good questions and the discipline to verify what a machine tells you. Curiosity has always mattered in this field, but in an AI-driven world, it is existential. The developers who keep learning will pull further and further ahead of those who don’t.
Education systems need to respond by becoming much faster and much more honest, because a four-year curriculum approved by a ministry cannot keep up with technology that reshapes itself every six months. Honest, because pretending students aren’t using AI, or worse, treating them as cheats when they do, prepares them for a workplace that no longer exists. The systems that will thrive are the ones that treat AI as a colleague to be managed, not a threat to be controlled. Estonia has always been good at that kind of pragmatism, and that is what we are trying to scale — in Nairobi, in Odesa, in Chișinău and wherever the model is needed next.
