Roman Yangarber

This postdoctoral opening sits in the Department of Digital Humanities, Language Technology at the University of Helsinki and is linked to two research directions: Know-AI and DALAI-FIN / Revita . Know-AI examines what language models know, when they know it, and how to make them more explainable and trustworthy for detecting and analyzing global threats such as disinformation in online media. DALAI-FIN / Revita focuses on AI support for second-language teaching and learning for adults, including migrants in Finland and elsewhere in Europe, with an emphasis on personalized instruction, learner modeling, learning analytics, and continual assessment of learner skills. The role is research-intensive and centered on language technology, large language models, and applied AI. The successful researcher will work on methods to evaluate, interpret, and improve the behavior of language models in threat-related settings, while also contributing to interdisciplinary collaboration across project teams and external partners. The position includes algorithm development, data analysis, and publication-oriented research in premier AI, ML, and NLP venues, as well as presenting work at conferences and workshops. The projects are described as internationally connected and collaborative, with a focus on cutting-edge work in interpretability, robustness, real-world safety, and personalized language-learning technologies.