Tseren Andzhukaev is the co-founder and CEO of NCSpeech, a company collecting AI training data and developing speech technology. In October 2026, the team introduced updated models for Kazakh, Russian and mixed-language speech. Recordings collected through cooperation with inDrive formed the basis of that development. (ncspeech.org, astanahub.com)
Who Tseren Andzhukaev is
Andzhukaev moved into entrepreneurship from research and development. His specialisms include machine learning, speech recognition, speech synthesis and voice biometrics. These areas are listed in his professional profile on NCSpeech’s website. (ncspeech.org)
Before establishing the company, he worked in private-sector research teams and published scientific papers. In his Digital Business interview, he also discussed working on Ayta AI, a project founded by inDrive founder Arsen Tomsky. Alongside Artyom Ivanov, he helped develop a product for people who stutter. (digitalbusiness.kz)
Ayta AI’s website identifies Andzhukaev as a research engineer. That experience concerned the practical use of speech technology with human speech that may differ from the standard examples available to a model during training. (ayta.ai)
How Tseren Andzhukaev built NCSpeech
Andzhukaev dates NCSpeech’s creation to March 2025. He identifies Nuraly Zhanbyrbayev, Dmitry Sandzhiev and Artyom Ivanov as fellow co-founders. The team brought together technical knowledge and experience working with corporate customers. (digitalbusiness.kz)
The startup initially offered ready-made speech solutions to large organisations. Customer discussions showed a demand for properly collected, cleaned, checked and labelled recordings. This became particularly clear in projects dealing with mixed Kazakh and Russian speech. Working through those requirements gradually led NCSpeech to develop data collection as a business line. (kz.kursiv.media)
In May 2026, Andzhukaev described the transition from engineering to running a business. Developers had to build the product while also selling it and establishing relationships with corporations. Sandzhiev’s arrival strengthened the team’s commercial capabilities. (digitalbusiness.kz)
Collecting training data
NCSpeech places tasks inside applications with large user audiences. A participant can record speech or complete another assignment in return for a payment. The company organises checks on the material so it can be used to train AI models. (kz.kursiv.media)
The founders described work with audio, photographs and video in May 2026. Preparing a dataset involves distributing assignments, checking quality and reviewing the content. The team used local linguists for language projects. Its customer receives material assembled for a particular task, language or environment. (digitalbusiness.kz)
In October, Andzhukaev told Kursiv about cooperation with inDrive and work in Southeast Asia. At Glovo’s Startup Campus in Barcelona, he proposed a potential delivery-platform use case: users could complete data-collection assignments while waiting for an order. (kz.kursiv.media)
The proposal formed part of discussions about possible cooperation. The programme gave the team access to Glovo executives and an opportunity to explore how its service might fit a large platform. For Andzhukaev, the visit continued the work of building corporate relationships outside Kazakhstan. (kz.kursiv.media)
NCSpeech Studio and language models
On 13 October 2026, Astana Hub reported an update to NCSpeech’s voice models. They support Kazakh, Russian and speech mixing the two languages. In his comments, Andzhukaev explained the need to accommodate speakers who change languages within a single sentence. (astanahub.com)
NCSpeech Studio offers transcription, text-to-speech, voice creation from a recording and integration with other products. Astana Hub described potential uses including meetings, interviews, lectures, subtitles and voice services. (astanahub.com)
The link between distributed recording collection and model development is direct: the cooperation with inDrive provides material used to improve understanding of natural speech. Application users participate in the recording process for a reward. This connects the company’s data-collection work to its own speech products. (astanahub.com)
More founder profiles from RAEM: Sanzhar Myrzagalym: who is the founder of Rette and Alibek Dostiyarov: who is the founder of Perceptis.
Funding, awards and plans
In May 2026, the team reported $110,000 raised from Antler and plans to seek approximately $3 million more. Andzhukaev linked the planned round to faster development and finding an investor able to contribute international contacts and expertise. (digitalbusiness.kz)
NCSpeech won the AI Future Builders category at alem.ai battle on 31 March 2026, receiving a KZT 10 million prize. In autumn, it was also recognised in the Market Shaper category at Astana Hub Battle during AI & Digital Bridge 2026. (ncspeech.org, digitalbusiness.kz, astanahub.com)
Timeline
- March 2025 — NCSpeech was founded. (digitalbusiness.kz)
- March 2026 — Won AI Future Builders at alem.ai battle. (ncspeech.org)
- October 2026 — Received the Market Shaper award at Astana Hub Battle. (astanahub.com)
- October 2026 — Andzhukaev participated in Glovo’s Startup Campus in Barcelona. (kz.kursiv.media)
- October 2026 — Introduced updated models for Kazakh, Russian and mixed speech. (astanahub.com)
Frequently asked questions
Who is Tseren Andzhukaev?
He is NCSpeech’s co-founder and CEO, with a background in machine learning and speech technology, including research and applied speech projects. (ncspeech.org, digitalbusiness.kz)
When was NCSpeech founded?
Andzhukaev dates its founding to March 2025. His fellow co-founders are Nuraly Zhanbyrbayev, Dmitry Sandzhiev and Artyom Ivanov. (digitalbusiness.kz)
What investment has the team disclosed?
In May 2026, the founders reported $110,000 from Antler. They also described a plan to raise approximately $3 million. (digitalbusiness.kz)
Which languages do the updated models support?
The October 2026 release covers Kazakh, Russian and mixed Kazakh-Russian speech. The capabilities are available through NCSpeech Studio. (astanahub.com)
