Research depth.
Production discipline.

A small nation. A remarkable scientific legacy.

Finnish scientists have helped build the foundations of artificial intelligence for more than half a century. In 1970, Seppo Linnainmaa developed reverse-mode automatic differentiation, a mathematical technique that would later become fundamental to training modern neural networks.

Teuvo Kohonen became one of the world’s most influential neural network researchers. His self-organising maps transformed how computers discover patterns, relationships and structures in complex data. At Helsinki University of Technology, he helped establish a research environment that educated generations of scientists in machine learning and computational intelligence.

Utopia Analytics emerged from this scientific tradition. Our doctoral researchers have roots in Kohonen’s research environment, connecting our company to one of Finland’s most distinguished schools of AI research.

Teaching machines to learn across languages

Co-founder Dr. Mari-Sanna Paukkeri pursued methods that could discover meaningful patterns in text without extensive language-specific rules or resources. Instead of building a new collection of dictionaries and rules for every language, her research explored how algorithms could discover relevant structures directly from data.

In 2008, Paukkeri and fellow researchers published the Likey method, demonstrating language-independent keyphrase extraction through statistical analysis. She developed this work further in her 2012 doctoral dissertation, Language- and Domain-Independent Text Mining.

The ambition was not simply to teach computers another language. It was to develop methods that could learn across languages.

Research brought into practical technology

In 2014, Utopia Analytics was founded to transform advanced machine learning research into practical technology. Our vision was straightforward: what if machines could learn how humans moderate content, instead of requiring people to write endless rules for every possible situation?

Our technology learns from human moderation decisions, recognises relevant patterns and adapts to each customer’s requirements. By analysing content in context, it addresses multiple languages, changing expressions, spelling variations and evolving moderation policies. This approach grew from a research tradition that predates the current wave of generative AI.

Finnish research. Patented innovation.

That work led to patented content moderation technology, including a United States Patent, granted in December 2022. Utopia Analytics founders Mari-Sanna Paukkeri and Tom Packalén are among its inventors—an important milestone in bringing Finnish academic research into practical technological innovation.

The underlying principle remains central to our philosophy: build intelligent systems that learn from data and human decisions, adapt to changing needs and reduce dependence on manually constructed rules.

Our heritage is our advantage

Research matters. Engineering matters. Reliability matters. These Finnish principles shape how we build technology: it should solve meaningful problems, work reliably and deliver measurable value. Our roots in scientific research and independent innovation continue through real-world applications today.

We did not enter artificial intelligence because it became fashionable. We built our company on the science behind it.