♟️ Chess, strategy, and the AI community — that’s how our very first chess tournament went! This Sunday, we gathered in Kyiv to play 7 rounds of rapid chess, compete for the win, and connect with the AI community.

👉 Vyacheslav Miekh, AI Product Owner at , took first place in the amateur category.
👉 Roman Yaremenko, Lead Software Engineer, came out on top among professional players.

A big thank you to the Chess Federation of Ukraine and Chess Club for their support and help in making the tournament happen. And to everyone who joined us — thank you for the games, the competitive spirit, and the great atmosphere. Looks like this was just the first move 😉

We gathered in Lviv for the Align, Forget, Repeat Meetup — a technical meetup focused on the behavior and memory of large language models 🔥

Together with Maria Korolyuk, Research Fellow at LASR Labs, and Viktoria Makovska, Research Engineer at Lapa LLM and PhD researcher at UCU, we explored how LLM behavior is shaped, why alignment does not emerge automatically, and how researchers work with the internal structure of models to make them safer and more controllable.

A big thank you to Maria and Viktoria for sharing their expertise and insights, and to everyone who joined us for the engaging discussions and thoughtful questions.

Data Science School 2026 — done 🔥

This year, 27 participants joined the school from Kyiv, Lviv, Kharkiv, Ternopil, Drohobych, Ivano-Frankivsk, and Chernivtsi. Together, they represented more than 10 educational institutions — from Lviv Polytechnic, KPI, and KhPI to KSE, UCU, and even the University of Pennsylvania.

Everyone came with different goals: some wanted to gain a deeper understanding of Data Science, others aimed to work on personal projects, build their portfolios, strengthen their practical skills, or connect with like-minded people. We believe that by the end of this week, everyone had taken a meaningful step closer to achieving their goals.

We're wrapping up the Handwritten to Data AI Challenge, which brought together more than 870 participants from Ukraine, Japan, and other countries, 198 teams, and over 6,800 Kaggle submissions around a shared goal: training AI to accurately recognize Ukrainian handwritten text.

The winners are:
🥇 1st place — Kostyantyn Zahorulko (Data Scientist, Mriia) and Oleksandr Dudniev (Data Scientist).
🥈 2nd place — Dmytro Kovalenko (Senior ML Engineer, Jaroona).
🥉 3rd place — Tetsuya Fujimura (Product Manager at Pluszero), a participant from Japan who chose to donate his prize to support Ukraine👏

Most importantly, the winning solution will not remain just a competition prototype. It is set to be integrated into public digital services, including upgrading the handwriting recognition model used in eDozvil, digitizing archives together with the State Archival Service of Ukraine, and supporting the automated assessment of homework and court documents.

UN Women AI School is a hands-on educational program designed for women who want to understand how artificial intelligence works and confidently apply AI in their work, projects, and everyday tasks. The program combines online learning, practical AI Labs, and real-world case studies led by industry experts.

We summed up the results of the Deep Tech Accelerator: Hardware Edition — and presented them at Demo Day.

Teams showcased what their journey over these past months has looked like: what they built, how their products evolved, and how their thinking has changed. In the expo zone, they also demonstrated how their startups work live.

The winner is TechNovator, a startup developing wireless power transfer technology for simultaneously charging devices at a distance and in motion. The team received $10,000 to further develop their product.