Taalai Djumabaev previously worked as a software developer at Deutsche Bank and Google before leaving the corporate world in 2023 to focus on educational projects. He started with a mentorship academy for IT professionals and later launched LUWO, an interactive platform for computer science lessons in schools. He invested $350,000 of his own money into building the platform and raised another $415,000 from investors.
In an interview for “100 Startup Stories from Central Eurasia”, a joint project by Digital Business and Astana Hub, Taalai talked about why he gave up a $500,000-a-year salary, how he got more than 100 schools using his platform in just a few months, and what kids really need to learn in the age of AI.
“At Google, I had everything I needed to live comfortably. But I wanted to see the difference my work was making”
– Taalai, could you tell us a bit about your background? What did your career look like before you started your own projects?
– I was born and raised in Kyrgyzstan. In 2005, I moved to St. Petersburg to study at the Faculty of Mathematics and Mechanics at St. Petersburg State University, enrolling in a five-year specialist degree program. I graduated in 2011. While I was studying, I spent two summers doing internships in Cambridge at the European Bioinformatics Institute (EMBL-EBI). After graduating, I joined JetBrains.
In May 2012, I moved from St. Petersburg to Moscow to work at Deutsche Bank. They were building a new strategic team to develop products capable of handling huge volumes of stock market data. I really got into high-load systems and spent the next ten-plus years working in that field. In 2016, I transferred to Deutsche Bank’s New York office in a similar role. I worked there for another three years and got my green card.
In 2019, I joined Google’s Insights & Opportunities team within its Ad Manager advertising platform. We worked with major clients like Spotify, Disney, and The New York Times, and handled up to a trillion events a day. We used machine learning to run simulations that showed clients how they could boost their advertising revenue by placing video ads in different spots on their websites.
– Why did you decide to leave Google?
– At Google, I was earning up to $500,000 a year. I had health insurance, opportunities to travel, and a perfect work-life balance. I had everything I needed to live comfortably. But I wanted to see what difference my work was actually making.
I had an Instagram blog where I shared bits of my everyday life, my work, workouts, and travels. One day, I asked my followers how I could help them. Around ten people got back to me, saying they wanted to learn from my experience and have me as their mentor. The very next day, I set up a company in the US called GrowthHungry Academy. We offered a year-long online program for developers, software testers, DevOps engineers, data scientists, and data engineers from around the world.
Students at the academy got a solid foundation in programming. They studied algorithms, discrete mathematics, and the principles of professional software development. Unlike other similar programs, all our mentors were working engineers at Google.
For just over three months, I worked on the academy alongside my job at Google. In September 2023, once the academy was bringing in enough money to cover its basic expenses, I left Google.
– How did you go from running a mentorship academy for adult developers to building an interactive platform for computer science lessons in schools?
– I kept seeing the same problem. An experienced engineer would join the academy, pay a lot of money, and get paired with some of the best mentors from Google. But we'd keep going back to things they should have learned at school. These were professionals who ended up spending thousands of dollars and countless hours filling gaps in their knowledge that should have been covered years earlier. Why wait until they're adults to address those gaps? It got me thinking about creating something for kids instead. In September 2025, we enrolled our final group of students at GrowthHungry, and I moved back to Kyrgyzstan from the US.
I started talking to principals and founders of private schools, and they all said the same thing: computer science education was a mess. The curriculum was outdated, good teachers were hard to find, and it wasn't always clear what kids were even being taught in class. A teacher might spend six months teaching Python and then quit. Halfway through the school year, a new teacher comes in who only knows HTML, for example, and the whole curriculum changes. So what a child learns ends up depending on whichever teacher happens to get hired.
At first, I wanted to open my own school. Three people I knew were even willing to fund its construction. But opening one school wouldn't change what was happening at all the others. So I started thinking about building a platform that different educational institutions could use.
“I invested $350,000 of my own money in development and raised another $415,000 from investors”
– You mentioned on social media that it took about 10 months to build the platform. When did you have the first prototype ready?
– We started development in December 2025. I hired one developer and did some of the coding myself, while my wife helped with the organizational side of things.
In May 2026, we ran our first pilot at Ilim, a private school. The pilot showed us there was real demand for the platform. After just one week of presentations at different schools, we had preliminary agreements to bring 1,700 students onto the platform.
– How did you fund the development? Did you use your own money or raise funding from investors?
– At first, we funded the project with our own money and resources from GrowthHungry. After our presentations went well in May 2026, I reached out to friends who were entrepreneurs and engineers I knew from Google and invited them to invest in the startup through a SAFE agreement. Basically, it means investors put money in now and get a stake in the company later, usually during the next funding round. We set a $5 million valuation cap for calculating their future equity stakes and raised our first $195,000 this way.
By July 2026, we had received requests from schools covering 17,000 students, and some schools had already paid deposits of up to 30%. We also closed two more SAFE deals, this time with a $12 million valuation cap. CS Angels, a business angel club based in Tashkent, invested $100,000.
Astana Hub Ventures, a venture capital fund, invested another $120,000 for a 1% stake in the company. For us, the investment is also valuable because of the connections it brings in Kazakhstan. Around ten local schools have already shown interest in the platform, and we're hoping the Astana Hub team can help us bring in more clients and grow our business in Kazakhstan.
In total, I put in around $350,000 of my own money and raised another $415,000 from investors. So far, we've spent about $500,000 of that on developing and launching the platform.
– Where is the startup based now, and how does your team work?
– Our parent company is registered in Delaware, in the US, and we have two subsidiaries in Kyrgyzstan and Uzbekistan. There are now more than 20 people on the team, including 10 developers, a sales department, a marketing team, and our own production studio. We hire people from all over the world. We've got team members from Belarus, Georgia, Russia, Kazakhstan, Kyrgyzstan, and Cyprus. Nobody's quite ready to relocate permanently to one country yet, but everyone's happy with the idea of spending the winter by the ocean. So around 10 of us are heading to Vietnam soon, where we'll stay until spring. We're planning to live together and work in what will basically be one big, non-stop hackathon.
“If Elon Musk launches a rocket to Mars tomorrow, kids will be learning about it in class the very next day”
– How does learning work on the LUWO platform?
– A teacher opens LUWO and finds a ready-made lesson with learning objectives, slides, and exercises. They don't have to create lesson materials from scratch every time.
During the lesson, the teacher's computer becomes a full-on control center. From there, they can display slides, interactive exercises, and quizzes on students' screens. The system also shows them in real time which students have switched to another browser tab and which ones are struggling with a particular task. A student might sit quietly, stuck on a problem, without raising their hand, but the teacher can still see they need help.
– Who develops the courses, and which educational standards do you follow?
– We base our curriculum on international computer science education standards.
We call it the Real-Time Curriculum because it evolves alongside the digital world. Say Elon Musk launches a rocket to Mars tomorrow. By the very next day, we could be using it as a real-world example in class. There's no need to wait for the next edition of a textbook to talk to kids about what's happening right now.
We also have a former computer science teacher on the team who now works in development. We regularly consult with external curriculum specialists, including experts from Harvard and MIT.
Every lesson goes through dozens of rounds of testing in real classrooms. Our team stays in touch with all the computer science teachers who use LUWO, and we constantly improve the platform based on their feedback and suggestions. We also use AI agents to analyze how students interact with the platform, which helps us make dozens of improvements to the lessons every day.
– How is your approach different from traditional computer science lessons at school or coding courses for kids?
– Coding courses teach kids how to write code. But will we even need traditional coders in 10 years? Our goal is to give every student a solid foundation in digital literacy and critical thinking.
First, we teach kids how to stay safe in today's digital world. They learn how to store passwords securely, spot online scams and deepfakes, deal with cyberbullying, practise good digital hygiene, and write professional emails.
Second, we focus on skills kids can use anywhere, rather than teaching them which buttons to click in a particular program. For example, when they work with spreadsheets, they learn how to analyze data, filter large datasets, and spot anomalies.
Third, we use programming and algorithms to help kids develop systems thinking. They learn how to break complex problems down into manageable steps. This also helps build critical thinking skills, which are especially important in the age of AI. People could easily start blindly trusting everything AI generates. In the future, the most valuable professionals will be those who can analyze an AI-generated answer and spot hallucinations or flaws in its logic.
“Schools help us improve the platform, and we lock in their price for five years”
– How does the platform make money?
– We deliberately set the initial price at just $20 per student per year. That barely covers our costs. We made a simple deal with schools: they help us test and improve the platform every day in real classrooms, and we lock in that price for them for five years.
– What challenges do you face when signing contracts with schools?
– Teachers need to understand how the platform will change the way they work. Some worry it might replace them, while others find it difficult to change how they're used to teaching. In our experience, those initial concerns often disappear after a couple of weeks of using the platform. Teachers see that they spend less time on routine tasks and have more time for their students.
Another challenge is getting schools to understand how the platform works. They're used to being offered either a standalone Python video course or an electronic gradebook. But with our platform, we have to demonstrate the whole concept of an interactive control center for running lessons in person.
– How many schools are using LUWO so far?
– We now have more than 100 schools using the platform across Kyrgyzstan, Uzbekistan, Kazakhstan, and Tajikistan. We've signed contracts covering 26,000 students, and $140,000 has already reached our company accounts. Many schools initially paid deposits of between 5% and 30% so they could test the platform in September, and they're now paying the remaining balances.
We're still getting class lists from schools every day. Some are running behind on renovations, while others are buying more computers. By the end of the calendar year, we're aiming to have at least 30,000 students using the platform, generate around $500,000 in revenue, and break even on operating costs.
Our more ambitious goal is to reach 50,000 students. Either way, we still have investment funds left, so the project is financially stable.
– What are LUWO’s main strategic goals for the next two years?
– We're preparing for a major expansion into Kazakhstan next school year. We also want to see whether schools outside Central Asia are interested in the platform. We already have meetings lined up with schools in Vietnam in November, along with a research trip to Singapore. We're also in talks with schools in Myanmar, Cambodia, Azerbaijan, Georgia, and Serbia, and we're exploring the markets in Thailand, Malaysia, and Indonesia. This winter, I'll be visiting schools in the Middle East, followed by Latin America in the spring.
Our main priority this school year is to test the platform in real classrooms and measure how much students are learning. Once we're operating in at least 10 countries and generating over $1 million in annual revenue, we'll consider raising a full seed round from global venture capital funds.