Article
Nitor has been accepted into an international research project tackling bottlenecks in industrial AI. The project includes 29 companies and research organisations from nine countries. Nitor will develop technical proof-of-concepts for various partner use cases within the data flow subset led by VTT.
Nitor is participating in the CLEAR research project, which is part of the international ITEA 4 cluster for software-intensive systems. The project's goal is to accelerate AI adoption so that players in industry, agriculture, and transport can lower operating costs, improve safety, and increase system reliability. Additionally, the project strengthens Europe's position at the forefront of industrial digitalisation.
“Without the means for systematic and scalable AI adoption, its value will be limited to scattered experiments. Nitor has developed a wide range of AI and machine learning solutions for its clients. We also have extensive experience with MLOps, which lays a strong foundation for developing operating models for AI deployment and management,” says Pasi Niemi, Nitor's Vice President of Research and Development.
Participation in the CLEAR project is a strategic investment for Nitor. Beyond the research itself, the aim is to further build expertise in an area that is rapidly becoming a critical competitive factor. Niemi expects the research project to bear fruit for Nitor's clients as well.
Our clients can rest assured that we are committed to being part of solving the obstacles ahead of large-scale AI adoption. Through this research, we hope to find a replicable model so that AI and its benefits reach our clients with increased speed and cost-efficiency.
Niemi coordinates the research Nitor carries out for the project's various subsets. In addition, Nitor has hired a full-time research assistant for the project. Amir Sultanbekov, M.Sc. in Systems and Operations Research, started in the role on June 1, 2025. Sultanbekov moved to Nitor from a role in business development and AI utilisation and specialises in machine learning, mathematical modelling and optimisation, and data analytics.
Multimodal data as a bottleneck for industrial AI
The CLEAR research project is divided into several research packages, each addressing the core challenge from a unique angle. Nitor works mainly within the multimodal pipelines work package led by VTT. This work package develops architecture and tools for AI deployment and management so that language and multimodal models leveraging complex data can operate reliably and predictably in continuous use. Nitor is set to build a flexible, security-focused AI platform that other participating organisations can use as part of their own solutions.
“Industry operations generate large volumes of so-called multimodal data, which flows from various sources and in various formats simultaneously. It is challenging to unify this kind of data in a way that supports decision-making and enables more efficient, safer, and more predictable production,” Amir Sultanbekov adds.
Multimodal data can refer to areas such as temperature sensor data, surveillance images and video, or text-based maintenance log entries. Large, complex volumes of data are difficult to utilise together in an impactful way.
The project is led by the ITEA 4 cluster for software-intensive systems, which is part of the international Eureka network for market-driven research and development projects. Eureka's cluster projects are long-term, large-scale, and internationally significant initiatives aimed at developing technologies important to Europe's competitiveness. In Finland, Eureka activities are managed by Business Finland, which also funds the projects carried out within the initiative.
