Artificial intelligence opens up new business opportunities, but not all users can take full advantage of its benefits. Most AI systems are designed for high-resource languages such as English, Spanish or Russian, which creates a serious language gap and deprives many, including Kazakhstanis, of access to advanced technologies in their native language needed to develop and improve lives in countries with low-resource languages. In Kazakhstan, the problem of the language gap in AI is being addressed by a scientific working group led by the Institute of Smart Systems and Artificial Intelligence at Nazarbayev University (ISSAI NU). They are working to create a large KAZ-LLM language model that will cover the key languages for Kazakhstanis: Kazakh, Russian and English, so that everyone can use digital technologies in their native language. This fundamental model will become the basis for the development of local services and products, as it will be published in the public domain. The digital operator Beeline Kazakhstan and its subsidiary IT company QazCode are partners in the creation of the national large language model KAZ-LLM. This is not the operator's first experience in
AI work. So, last year Beeline launched and made publicly available to all developers the Kaz-RoBERTA-conversational model, which is actively used to serve subscribers on digital platforms. It has been downloaded more than three thousand times on the Hugging Face platform. Such initiatives are especially relevant against the backdrop of the importance of digitalization, which the country’s leadership emphasizes. President of Kazakhstan Kassym-Jomart Tokayev emphasized the importance of strengthening the country’s digital infrastructure and expressed his readiness to personally oversee the creation of an AI system in Kazakhstan. “A new era is unfolding before our eyes. The impact of AI technologies is as revolutionary as the discovery of electricity and the Internet. And the development of AI must outpace the needs of IT developers. First of all, it is necessary to increase our computing power,” the President said, emphasizing the importance of the development of domestic technologies. Thanks to Beeline Kazakhstan and QazCode, the project to create KAZ-LLM, supervised by ISSAI, is provided with the necessary infrastructure. The company provided cloud computing power of 8 DGX H100, which made it possible to significantly increase the volume of training data and model training capabilities. TO
the company also provided the collected open data, and QazCode data scientists joined a joint working group to train the model. "KAZ-LLM will be able to create content in the languages most relevant to Kazakhstan: Kazakh, Russian and English. The model will play a critical role in preserving the nation's cultural heritage and will cover historical context, specialized areas and conversational data representative of Kazakhstan. By adapting generative AI to local needs, KAZ-LLM will demonstrate how national projects can bridge language gaps and contribute to the global AI innovation landscape. Most importantly, the project KAZ-LLM contributes to the creation of cutting-edge generative AI specialists. Through a hands-on approach to data preparation, training and model implementation, Kazakhstan is supporting a new wave of cutting-edge scientific talent capable of creating generative AI models and tools,” ISSAI NU commented. This contribution to the development of the national large language model KAZ-LLM reflects the strategy of the digital operator Beeline, as well as the desire to bridge the language gap in AI technologies. From this point
The operator signed a memorandum with the Supercomputer Center of Barcelona, which specializes in the development of AI in different language groups, and also announced the creation of the first GPU cloud in Central Asia for the development of AI products based on NVIDIA technology. Today, the company’s portfolio includes many successful AI projects for business: video analytics for sales, computer vision, video surveillance in production, marketing solutions and other products. "Our accumulated experience, knowledge and cooperation with the Barcelona Supercomputer Center allow us to focus on three key areas. Firstly, the creation of a domestic supercomputer for processing large volumes of data and providing access to this computing power to Nazarbayev University specialists for training models. Secondly, we continue to develop the Kazakh language model Kaz-LLM in order to integrate it into the digital space. And finally, we pay special attention to the development of professionals in the field Data Science, so that they can successfully work on complex projects and compete in the international IT arena,” said Alexey Sharavar, CEO of QazCode. Development of large language models in such complex projects
Such as KAZ-LLM plays a key role in the formation and training of strong personnel in the field of Data Science for Kazakhstan. These specialists not only gain unique experience working with advanced technologies, but also lay the foundation for further progress in the national AI industry. The joint efforts of Beeline Kazakhstan, QazCode, Nazarbayev University, a consortium of the country's leading universities, as well as the Ministry of Digital Development, Innovation and Aerospace Industry of the Republic of Kazakhstan and the Ministry of Science and Higher Education of the Republic of Kazakhstan not only promote AI technologies in the Kazakh language, but also contribute to the further development of the country's digital economy. The most important things about events in Kazakhstan and the world
