Citizens of Silopolis: Jussi Karlgren, Principal AI Scientist at SiloGen

jussi karlgen profile picture

Jussi Karlgren is a mathematically inclined computational linguist and Principal AI Scientist at Silo AI. He holds a PhD in computational linguistics from Stockholm University and the title of docent (associate professor) of language technology at Helsinki University. Jussi joined Silo AI from Spotify, where he worked on personalization and search on podcasts. At Silo AI, Jussi is a part of the SiloGen team, and in his work, he focuses on quality assessment of generative language models and on new ways to apply generative AI to real-world problems. 

Jussi has a research background in computational analysis of genre and stylistics, distributional semantics, and information retrieval. He is known for having first formulated the notion of a recommender system. Jussi worked as a researcher at SICS (now RISE ICT) for 20 years and co-founded Gavagai, a text analysis company, where he worked for ten years. Most recently, he was a researcher at Spotify. Additionally, he has held various research-related roles in research groups in academic and industrial organizations, such as Stanford University, Xerox PARC, Recorded Future, and KTH Royal Institute of Technology, and advisory roles to government agencies, NGOs, and technology. He participates in academic activities by organizing international research workshops and symposia and supervising MSc and PhD students at Stockholm University, KTH, and the UK. 

Jussi's journey in AI: From knowledge-based systems to generative models 

When Jussi first came across a class on "AI programming" at the computer science department of Stockholm University, where he was studying linguistics and mathematics, he was thrilled. The year was 1987, and AI at that time mainly consisted of knowledge-based systems with less emphasis on machine learning. Fortunately, that year was also when the book on Parallel Distributed Programming was published. This marked the rebirth of connectionist or neural processing frameworks, which is what has given us today's generative models. Jussi attended a summer school on neural models, and while it took a while until he was in a position to build connectionist models himself, he was hooked already.  

The books that sparked Jussi's interest in AI and set him on his career path: The Art of Prolog by Leon S. Sterling, Ehud Y. Shapiro, Structure and Interpretation of Computer Programs by Julie Sussman, Parallel Distributed Processing by Rumelhart David E., McClelland James L.

In the meantime, Jussi worked on computational stylistics, information retrieval, media analysis, and other statistical approaches to large-scale processing of text, first as a researcher at SICS, a research institute in Stockholm, as a visiting scholar at Stanford University, and as an entrepreneur at Gavagai which he founded with his colleague Magnus Sahlgren and where they soon were joined by Fredrik Espinoza, both of whom also are at Silo AI today. Jussi's background and experiences made it only natural for him to join the Silo AI team.

Bridging the gap between research and the real world: Jussi's work at Silo AI

Jussi is interested in both ends of the processing pipeline. He is interested in how human communicative practice is represented and how a system that is intended to communicate with humans can be designed to fit the expectations of its users. The former is a research-oriented challenge, while the latter is about evaluating a system as a tool. In his work at Silo AI, Jussi gets to combine both aspects, and that is actually something he mentions as his favorite part of his work – seeing his work in practice as the research results move into production swiftly.

"There's always something new to do and learn. What we do has significant effects outside the laboratory. I have worked with research in many different contexts, and it is hugely exciting to be involved in an area where research results move into production this rapidly!"

In addition to his role, when describing his favorite parts in his work at Silo AI, Jussi mentions the unique community and strong value base that's so well-aligned with his personal values: 

"My own favorite personal value in life is helpfulness, and this cuts across all Silo AI values: Keep Learning, Ask Why, Build bonds, and Be good – being helpful is connected to all of them!"

Outside work

Outside of work, you will most often find Jussi near water, on a boat, or on a shore, usually in the Stockholm archipelago. When asked about his favorite activities, Jussi cites the Rat from “The Wind in the Willows”: 

"Believe me, my young friend, there is nothing - absolutely nothing - half so much worth doing as simply messing about in boats. In or out of them, it doesn't matter."

To learn more about Jussi, his research, and background, visit his website here.

Would you like to work with pioneering professionals like Jussi? Join our team! 

We're constantly growing our team, check out our open positions here:


No items found.

Want to discuss how Silo AI could help your organization?

Get in touch with our AI experts.
Mia Jokiluhta
Content Marketing Lead
Silo AI

Share on Social
Subscribe to our newsletter

Join the 5000+ subscribers who read the Silo AI monthly newsletter to be among the first to hear about the latest insights, articles, podcast episodes, webinars, and more.

What to read next

Ready to level up your AI capabilities?

Succeeding in AI requires a commitment to long-term product development. Let’s start today.