Published on: 2026-08-18
Source: Official website of the State –
An important disclaimer is at the bottom of this article.
Artificial intelligence is becoming part of companies’ work, but to use it effectively, it is not enough just to connect a neural network. It is important to teach it to correctly work with corporate documents and find the necessary information in them. This is precisely the task solved by the project of GUU student Nikita Kolotushkin, who became the winner of the “UMNIK” contest by the Innovation Assistance Fund.
Nikita is a student at the Institute of Information Systems, a technician at the Laboratory of Digital and Intelligent Mobility, and a member of the Student Design Bureau of the State University of Management “Innovative Solutions.” Winning the competition earned him a grant of 500,000 rubles to implement his own project.
The “UMNIK” competition is held by the Foundation for Assistance to Innovations in cooperation with the Ministry of Economic Development of Russia within the federal project “Technologies” of the national project “Effective and Competitive Economy.” Its goal is to support young researchers and help promising scientific and technological developments take the next step towards practical application.
How to teach artificial intelligence to work better with documents?
Nikita’s project is dedicated to one of the important problems of modern AI systems — high-quality preparation of the information that the neural network works with.
The full title of the development sounds complicated: “Development of a software complex for automated control of machine readability and preparation of industrial text data for search-augmented generative systems based on structural-linguistic metrics.” But behind this scientific formulation lies a perfectly understandable task.
Imagine that a company wants to create an AI assistant that can answer employees’ questions using thousands of internal instructions, regulations, reports, and other documents. If you simply feed the neural network a huge array of unstructured texts, it will be difficult for it to quickly find exactly the information needed for an accurate response.
Nikita’s development helps prepare such documents for working with artificial intelligence: cleaning source data, extracting necessary information, splitting large texts into small meaningful fragments, and adding information that will help the system navigate the knowledge base.
After such processing, AI can find relevant fragments faster and use them when generating a response. This approach is called RAG (Retrieval-Augmented Generation). Simply put, the neural network first searches for the necessary information in the prepared knowledge base, and then generates a response based on it.
Today, companies are increasingly integrating artificial intelligence into their workflows. Therefore, the quality of corporate data preparation becomes one of the key conditions for the effective operation of such systems. Nikita’s project is precisely aimed at making this process more accurate and automated.
From student competitions to our own development
Winning in “UMNIK” became a continuation of Nikita’s active project work. In 2025, he became the winner of the “Doing” track of the All-Russian student project “Your Move,” as well as won the hackathon “Urban Development Technologies.”
In 2026, Nikita also joined the All-Russian project of the Ministry of Education and Science of Russia “Tech Mentors”, one of the tasks of which is to engage children and youth in engineering and technical creativity.
In total, in 2026, authors of 500 scientific and technological projects received grant support from the Foundation for Assistance to Innovations. Participants’ applications were reviewed at 17 venues, and winners came from 66 regions of Russia.
Now the winners have to move from the idea to research and practical implementation of their developments. In the future, the results of this work may become the basis for creating their own small technology companies.
Congratulations to Nikita on the victory and we wish success in the project implementation!
Subscribe to the Telegram channel “Our GUU” Publication date: 08/18/2026
Please note; This information is raw content obtained directly from the information source. It represents an accurate report of what the source claims and does not necessarily reflect the position of MIL-OSI or its clients.