Skoltech Global

🎉 Celebrating the fifth InteRussia fellowship at Skoltech

Skoltech, in cooperation with the Gorchakov Fund and Rossotrudnichestvo, has officially inaugurated the new cycle of the InteRussia fellowship program, marking the fifth joint initiative over the past three years.

The young specialists will explore Russia's scientific and technological advancements, participate in the Skoltech Innovation Workshop, and exchange expertise with peers from across the globe. This year, the program brings together 15 talented researchers and specialists from 13 countries:

🇧🇦 Bosnia and Herzegovina
🇮🇩 Indonesia
🇮🇷 Iran
🇮🇶 Iraq
🇯🇴 Jordan
🇰🇪 Kenya
🇲🇾 Malaysia
🇵🇪 Peru
🇷🇸 Serbia
🇿🇦 South Africa
🇹🇳 Tunisia
🇹🇲 Turkmenistan
🇻🇳 Viet Nam

Alongside an intensive academic agenda, the fellows will enjoy a vibrant cultural program, including a performance of The Nutcracker on the historic stage of the Bolshoi Theatre, a visit to the State Tretyakov Gallery, and a guided tour of Moscow.

The InteRussia fellowship program for foreign specialists in STEM is implemented by the Mezhdunarodniki Autonomous Non-Profit Organisation in cooperation with the Gorchakov Fund and Skoltech. The program is implemented as part of Rossotrudnichestvo’s flagship New Generation initiative with grant support from the Presidential Grants Foundation.

Learn more on the InteRussia website.

🧠 How to keep language models honest?

Large language models are great at almost everything, except one thing: they excel at fabricating convincing nonsense, even when connected to knowledge bases via RAG. Traditional methods for tackling hallucinations usually demand either tons of labeled data or heavy computations involving repeated answer generation.

Researchers from Skoltech and Sber's Center for Applied AI have found a much more elegant solution — a new method called TOHA (TOpology-based HAllucination detector).

🔹 Instead of putting the model through heavy extra workloads, the approach looks "under the hood" to analyze the topological structure of its attention maps using the MTop-Div metric. If the response graph diverges from the context, the algorithm instantly flags the discrepancy.
🔹 The method requires no training of additional classifiers and uses minimal labeled data.
🔹 In terms of accuracy, it easily competes with resource-heavy alternatives like SelfCheckGPT.

The tool is already integrated into Sber's open-source SIRIN library, ready to protect corporate knowledge bases from AI fabrications.

The research has been accepted to the main track of the A* ACL 2026 conference. Dive into the details and author comments in our full release.

🧲 Skoltech researchers assess the behavior of pultruded thermoplastic composites at elevated temperatures

Researchers from Skoltech (part of the VEB.RF Group), together with colleagues from Xi'an Jiaotong University and the Blagonravov Institute of Mechanical Engineering of the Russian Academy of Sciences, conducted a systematic study of the flexural properties of pultruded glass fiber composites based on three different thermoplastic matrices: polyphenylene sulfide (PPS), polyamide-6 (PA6), and polypropylene (PP).

The results show that a composite's behavior under heating is determined primarily by the polymer matrix rather than the reinforcement. The findings will help engineers select materials for components operating at elevated temperatures in aviation, automotive, energy, and other industries. The research was supported by a Russian Science Foundation grant.

What the study revealed 👇

🔹 in our cards;
🔹 in the detailed article on the Skoltech website;
🔹 in the publication in Next Materials.

🔔 Class is in session! The traditional bell has rung, officially kicking off the new academic year at Skoltech.

🤝 Skoltech President, RAS Academician Yulia Gorbunova personally welcomed the new master's and PhD students, wishing them courage in their experiments, perseverance in their research, and dedication to their work.

⚡️ The grand opening of the Innovation Workshop took place at Skoltech. Viktor Vekselberg, Deputy Chairman of the Skoltech Board of Trustees, delivered a welcoming speech.

🚀 Skoltech at the HICOOL Global Entrepreneur Summit in Beijing!

Skoltech is participating in the HICOOL Global Entrepreneur Summit, taking place at the China International Exhibition Center.

At our booth, visitors can learn all about admissions and chat with Skoltech alumnus Evgeny Shilov, co-founder of Labadvance, one of Skoltech's most successful startups.

About Labadvance:
The company develops microfluidic chip technology. Powered by artificial intelligence, these smart chips with micro-channel networks replace bulky equipment, making chemical synthesis fast, precise, and compact. The technology is in high demand across pharmaceuticals, petrochemistry, advanced materials manufacturing, and scientific research.

Skoltech is part of the VEB.RF Group.

🔬 Single-walled carbon nanotubes are packed with potential, but working with them has always been difficult and costly. To get a tiny conductive pattern (just 100 micrometers in size), scientists traditionally had to deposit a solid film and etch away the excess, losing up to 90% of the expensive material to waste!

Researchers from Skoltech and their colleagues have found an ingenious solution to this problem, publishing their findings in the journal Light: Advanced Manufacturing.

«The key insight comes from aerosol science: carbon nanotubes are so light that their inertia is negligible — they just follow the carrier gas. By pairing that principle with a nitrocellulose membrane pressed with a hot laser stencil, we closed off the pores we don't need. In effect, we let the gas flow itself draw the pattern, depositing nanotubes only where they belong. Nothing is etched away, and the template is fully reusable,» commented one of the study's authors, RAS Professor Albert Nasibulin, who heads the Skoltech Photonics Center.

Thanks to this method, the team managed to completely avoid toxic solvents and preserve the high quality of the material.

«The new method is highly scalable, opening up promising possibilities in optics, electronics, and sensorics,» emphasized Nikita Raginov, a Skoltech research intern and the lead author of the study.

To demonstrate the versatility of the approach, the research team created devices of two distinct classes. The first is a mechanical strain sensor in a W-shape. Its sensitivity proved to be at least three times greater than analogs, and tests over 3,000 load-unload cycles proved its absolute reliability. Such sensors can be embedded anywhere from bridge pillars to aircraft wings. The second device is a stretchable lens for manipulating terahertz radiation, which is critically important for future 6G networks.

Beyond terahertz optics and high-performance strain sensors, the researchers foresee potential applications in transparent conductors, bioelectrodes, thermoelectric modules, and heating grids.

The structured carbon nanotube samples described in the press release were modeled at ITMO University and the Harbin Institute of Technology, manufactured at Skoltech, and tested at MIPT and the Prokhorov General Physics Institute of the RAS.

Skoltech is a VEB.RF group institution.

🇿🇦 Today Skoltech hosted an official visit by a delegation from South Africa’s National Advisory Council on Innovation (NACI), led by Council Chairperson and ST&I Business Forum Founder and CEO Tilson Manyoni.

The guests were welcomed by Skoltech President, RAS Academician Yulia Gorbunova. The discussions focused on the Science and Innovation Park (SIP) project, technological cooperation initiatives within the BRICS, and prospects for bilateral partnership.

During the meeting, participants were introduced to the institute's research and educational potential, as well as the architectural concept of Skoltech’s unique campus.

The program continued with a guided tour of Skoltech's laboratories, where the international colleagues were shown the latest developments in engineering solutions, materials science, and digital technologies.

Skoltech is a part of the VEB.RF Group.