JST Healthcare Solutions is developing Finnish health technology with the help of LUMI AI Factory
4 mins read
The development of health technology in Finland is currently being driven in particular by a shortage of healthcare personnel, an aging population, and rapidly advancing artificial intelligence (AI) technology. New digital methods, tools, and AI-assisted solutions are constantly being developed to identify and address individual care needs.
JST Healthcare Solutions, a health tech company founded in 2020, leverages LUMI AI Factory’s services, including the LUMI supercomputer, to develop AI-assisted symptom assessment tools for both telephone services and other digital channels. The company’s MDR Class IIa-certified Smart Symptom Check software is widely used by healthcare professionals and patients to support care needs assessments and the collection of preliminary information.
The company’s MDR Class IIa-certified Smart Symptom Check software, classified as a medium-risk medical device under the European Union Medical Device Regulation (EU) 2017/745 and subject to independent conformity assessment by a Notified Body, is widely used by healthcare professionals and patients to support care needs assessments and the collection of preliminary information.
The software collects patient symptoms in a standardized manner and enables automated patient guidance and data-driven decision-making. The solution can be integrated into digital services or used to support the assessment of care needs over the phone.
The software is based on an extensive symptom dictionary compiled over the years and the analytics derived from it. As the company has grown, AI-based product development has also been incorporated. The goal is to accelerate development by leveraging the latest technology and the computing power provided by the LUMI supercomputer.
Boost from the LUMI AI Factory
In the fall of 2025, the LUMI AI Factory brought together around ten companies in the health technology sector into a focus group to develop concrete use cases for leveraging the LUMI AI Factory’s services. The event was aimed at companies that had something to contribute to the joint development effort.
The focus group’s goal was to refine bold, scalable, and impactful AI ideas in collaboration with leading experts in the field. At the same time, the aim was to gather user experiences to support service development and to raise awareness of the opportunities offered by LUMI AI Factory.
Participants were asked to propose a pilot-level use case or a development need that could be moved toward concrete implementation using the resources of the LUMI AI Factory.
When JST was given the opportunity to test the LUMI AI Factory’s services in practice, the company began to consider how it could more effectively advance AI development. As a result, the company decided to take advantage of the LUMI supercomputer’s free Try&Buy testing option for six months.
Finnish symptom vocabularies, including dialects, as the data foundation
JST decided to focus its LUMI testing on developing its own language model, as the language model it had been using did not recognise the Finnish symptom vocabulary—including its various dialects and expressions—with sufficient comprehensiveness.
– We have built an extensive thesaurus of synonyms ourselves, which our software utilises. So we already have a wealth of data at our disposal, explains Johannes Kallio, JST’s Chief Medical Officer and founder.
The company also possessed the technical expertise needed to carry out the project. Application developer Juha Kervinen has long utilized high-performance computing in software development, so the implementation of LUMI went smoothly from the administrator’s perspective.
Although LUMI AI Factory offers its users training and implementation support, the actual development work is the responsibility of the companies themselves. In the day-to-day operations of a startup, however, client work must be handled alongside product development, which is why JST’s active testing on LUMI was done during the quieter summer period.
The company was given sufficient time to carry out the necessary test runs. The test results proved to be significant for the product’s further development.
– The LUMI supercomputer is extremely well documented, and as a former CSC employee, I was able to easily harness it for the development of JST’s RAG AI model. We set out to secure computing power, and we succeeded. We saw particularly significant benefits in July 2026, when the development project progressed rapidly, says application developer Juha Kervinen, who serves as JST’s LUMI administrator.
Promising results
JST has been extremely satisfied with its collaboration with LUMI AI Factory.
– It’s great that AI models can be trained in Finland efficiently and to a high standard to meet healthcare needs, says Kallio.
The company views LUMI as an important tool for the future as well, even though strict healthcare regulations affect the timelines for implementing new solutions.
– Regulation in our industry is extremely strict. Our further development is largely determined by the timeline for bringing the models into production in accordance with the requirements of the MDR and AI regulations. This naturally also affects our goals and the prioritization of development projects. However, we believe we will continue to utilize LUMI in the future, Kallio says.
Author: Tiina Leiponen, CSC – IT Center for Science
Image on top: Adobe Stock