Celsus story: from the first prototype to a leader in the medical AI market

Celsus is one of the leading domestic clinical decision support systems, capable of processing millions of studies and reducing the workload for medical specialists. This interview is about how the idea was born, the challenges we overcame, and the future prospects for medical AI.

KAMA FLOW spoke with Nikita Nikolaev, Marketing Director and Co-founder of Celsus.

• Could you please tell us more about what Celsus does?

Celsus is a clinical decision support system based on neural networks, designed to analyze digital medical images—X-rays, CT scans, MRIs, and more. It detects objects and automatically interprets results, helping radiologists prioritize case flow, identify pathologies, and optimize their workflow.

• How did the idea for the startup come about, and who initiated it? Why medical technology in particular?

The idea to create a company specializing in AI technologies to help doctors read and interpret X-rays and CT scans appeared in 2017. Artyom Kapninsky, Celsus’s Commercial Director and Co-founder, discussed the possibilities of artificial intelligence with me. Around that time, international competitions for algorithms capable of detecting cancer pathologies in medical images were just beginning.

Our team decided to try their hand at mammography. The social impact here is enormous: early diagnosis of breast cancer allows clinicians to intervene when the condition can be fully cured. Artificial intelligence turned out to be the optimal solution.

In autumn 2017, we began actively developing the first version of our AI system. At that moment, our team included the CTO, Evgeny Nikitin, and the General Director and founder, Stanislav Chernin—four people in total. By early 2018, we already had a full-fledged ML department in St. Petersburg and a product team in Kaluga. Our first medical partners were oncology clinics in Kaluga and Tambov. By November 2018, we had an MVP version of the product, which showed promising results.

• What resources did you have at the start?

Creating the MVP took about a thousand hours of work and up to five million rubles of investment. But that was just the first step: by the time we participated in the "Moscow Experiment on the Use of AI in Healthcare" (mosmed.ai), the team had grown to over 20 employees and over $1 million had been invested in technology development. At the same time, we were working on medical device registration with Roszdravnadzor, and expanding the product line.

From day one, the company worked in two cities. The St. Petersburg office was responsible for technical development, and the Kaluga team focused on commercial growth, interaction with the medical community, and financing.

• What technical challenges did you face?

In the late 2010s, the medical AI market in Russia was practically non-existent, so integrating the product into real clinical practice was a serious challenge. Several issues had to be addressed at once:

• Integrating AI into the systems doctors were already using (PACS, RIS)


• Organizing monitoring of research flows


• Developing methods of clinical evaluation


• Scaling and increasing efficiency in the cloud

This work laid the foundation for future company growth.

• Who was your first client?

The first pilots were launched at regional oncology clinics. Gradually, the project was scaled up, and the technology began to be used in various healthcare institutions—from small centers to large regional healthcare systems.

• Why did you seek investment, and how did you connect with KAMA FLOW?

To accelerate growth, we needed not just funding but also a partner that could help us navigate regulatory barriers and scale our product. This is the support that Celsus received from KAMA FLOW.

• How did KAMA FLOW help with your growth?

After the investment round, the company received its first registration certificate for an AI medical device at the highest risk class. This made it possible to fully commercialize our healthcare solutions.

By 2022, our product line had grown from two to four solutions. Celsus received the European CE Mark and ISO 13485 certification. Most importantly, our technology was being applied in real practice: in 2020, we processed tens of thousands of studies; by 2022, we were already processing over a million.

• Where are your technologies used today?

Celsus solutions are implemented in healthcare organizations around the country—from individual diagnostic centers to centralized regional systems. In 2023, our technology was used in 19 regions, representing about 45% of the available market; by mid-2025, our systems are already in use in most Russian regions.

Celsus also partners with medical equipment manufacturers: clinics can acquire devices with embedded AI.

In some areas, Russian AI solutions are on par with, or even outperform, foreign counterparts—such as in X-ray and fluorography analysis. The Celsus system can automatically interpret up to 75% of images without significant pathologies, reducing radiologists' workloads by 70%.

• What prospects do you see now?

The company plans to expand its presence both in Russia and abroad. Our priorities include adapting solutions for different healthcare standards, expanding our product lineup, and developing new diagnostic areas.