Selected work.

I have been part of building and leading Experimentorium since 2016, beginning before the museum opened in Tbilisi. It had exhibits, visitors, and a team. What it did not have was coherence — between what it was, what it looked like, and how it operated.

Like many cultural institutions outside major cultural and funding centers, it had been built piece by piece, by hand. Over time, the pieces had stopped fitting together.

I also saw something larger: small institutions rarely get rebuilt by international consultants or design studios. If an institution is going to become modern, someone inside has to learn how to give it form.

Over more than a decade, I rebuilt Experimentorium from the inside — as a public identity, an educational space, and an operating system.

I developed a new bilingual identity in Georgian and English, designed to meet the visual and communicative standards of contemporary science centres. I built the museum's website from scratch, reshaped its public language, and created a clearer visual system for social, print, and onsite communication.

I also worked on the deeper structure: educational programs based on inquiry rather than passive display, internal workflows for a small team, and AI-supported systems for content, scheduling, visitor communication, and message funnels.

The goal was never only to make the museum look better. It was to help it think, speak, and operate as one coherent institution.

Experimentorium began to read not as a local project, but as a modern educational institution — to visitors, partners, and its own team.

It became the first proof of my method: that small cultural and educational institutions can be rebuilt from within when vision, structure, and execution are held together.

Everything I now bring to other institutions, I learned by building this one first.

Programs for girls in science already existed. Many of them began as local initiatives, built around an enthusiastic founder, her energy, her network, and her city. They could be powerful, but they were often difficult to transfer. When the founder moved on, the structure often weakened. When the program left its original context, something essential could be lost.

I saw a second problem underneath the first. Many programs were built around the message "girls belong in STEM" — a message addressed to a perceived absence. But twelve-year-old girls do not respond to statistics about their underrepresentation. They respond to people. They respond to stories. They respond to becoming someone.

The gap was not motivational. It was structural.

There was no clear education model for girls aged 10–14 that combined identity formation through the real lives of women in science with a system designed from day one to travel — to other cities, other countries, and other cultures — without losing what made it work.

Girls Who Change the World is designed as more than a one-off programme: a repeatable education system that can later travel through trained local partners.

It is built across four seasons — Science, Technology, Medicine, and Planet — each centered on women who saw what others could not. Each cohort includes twelve girls, a scale small enough for identity formation and strong enough for belonging.

The model includes a long-term progression: participant, observer, mentor, guide, and lifetime member. A girl who enters at ten can return at twenty not as a former student, but as someone who shapes the program for the next generation.

I designed the full institutional architecture: the educational model and seasonal structure, the cohort and progression logic, and the complete identity system — from language and rituals to brand, member kits, badges, and naming conventions.

The programme is being developed for future licensing to science museums, schools and education organisations that want to run GWCTW locally.

The system is documented and designed to be reproducible without depending on its founder.

Girls Who Change the World now exists as a working architecture, being prepared to move beyond Tbilisi after the Founding Cohort. The Founding Cohort opens in Autumn 2026, with the international partner framework built into the model from year one.

The design question is whether a programme for girls can travel across contexts without becoming generic.

It is also one of the clearest examples of how I work as an Education Systems Architect: I build the structures movements need in order to exist beyond their founder.

The public conversation about AI in education had collapsed into two flat positions: ban it, because children will cheat, or embrace it, because this is the future.

Both sounded responsible. Neither was enough.

Underneath both positions was a quieter problem. After more than ten years of watching children learn inside a science museum, I had seen something most policy conversations missed:

AI did not break education. It revealed where education was already weak.

A child who has been taught to think does not lose that capacity the moment a chatbot appears. A child who has only been trained to produce the right answer has very little to defend.

The crisis was not the machine. It was that education, in many of its forms, had become too focused on the production of artifacts — homework, essays, answers, presentations — and not focused enough on the development of thinking.

What was missing was not another warning about AI, and not another celebration of it. What was missing was a coherent position educators could stand on, and a practical method they could teach from.

The Human Intelligence Method is a manifesto and practical framework for educators teaching in the age of artificial intelligence.

It does not argue for or against AI. It argues for a redesign of learning so that a child's thinking remains visible, defensible, and truly their own.

The work has three layers.

First, a founding manifesto structured in six parts — from the wrong conversation about AI, to the Five Principles of the Method, to a redesigned Marie Curie assignment, to the Twenty-Minute Practice that can fit inside a single class period.

Second, a public infrastructure for the method: a dedicated website, a manifesto distribution funnel, a one-page classroom practice, and an editorial voice that allows the idea to live, spread, and develop over time.

Third, a Founding Lab of 8–10 educators working with the method in real educational settings, using it on their own assignments, with their own students, and helping refine it through practice.

Everything is designed for one purpose: to give educators a clear position to stand on, and a working method to act from.

The Human Intelligence Method now exists as a public framework for a problem many schools are facing but few can yet name clearly.

Educators, school leaders, and curriculum designers can read it and apply it. The first Founding Lab is being prepared for a cohort of 8–10 educators.

Of all my work, this is the layer where the position itself lives — not how an institution looks, not only how it operates, but what education is still responsible for when answers become instant.

It is also one of the clearest expressions of my work as an Education Systems Architect: I build systems that protect what must remain human inside education.