Pepa Atanasova

The University of Copenhagen’s Department of Computer Science and Pioneer Centre for AI invite applications for a fully funded PhD fellowship in Trustworthy Natural Language Processing (NLP), starting March 2026 or as soon as possible thereafter. The position is supervised by Assistant Professor Pepa Atanasova and co-supervised by Professor Isabelle Augenstein, both leading researchers in the CopeNLU group. The research will focus on developing innovative methods for trustworthy NLP models, including mechanistic interpretability, explainability, reasoning, factuality, safety mechanisms, and practical applications of trustworthy AI. The specific research direction will be tailored in collaboration with the successful candidate to align with their interests and the group’s goals. Applicants should have a Master’s degree in Computer Science or a related field, strong English skills, and ideally experience in ML or NLP. The fellowship offers a vibrant, international research environment, with opportunities for collaboration, teaching, and knowledge dissemination. The position is fully funded for three years, with competitive salary and social benefits. The application deadline is 31 October 2025. To apply, submit all required documents via the official application portal. The University of Copenhagen is ranked among the top institutions for NLP research and offers an excellent quality of life in Denmark, one of the world’s happiest countries. The CopeNLU group is known for its collaborative and inclusive atmosphere, supporting diverse research interests in NLP and AI.

University of Copenhagen is recruiting a postdoctoral researcher for the project DeceptionWeb: Foundations for Emergent Deception in Human-AI Interaction , a collaboration led by Pepa Atanasova (Department of Computer Science / DIKU), Ingo Zettler (Copenhagen Center for Social Data Science, SODAS, and Department of Psychology), and Egor Starkov (Department of Economics). The postdoc will work at the interface of machine learning , behavioural psychology , natural language processing , LLM security , and explainable AI . The project studies how deceptive interactions can emerge in human-AI systems, how they can be detected, and how they may be mitigated. The role includes contributing to controlled human-AI interaction experiments, building computational evaluation infrastructure, and developing interpretability-based detection methods for a joint oversight framework. The successful candidate will join the CopeNLU group and SODAS at the University of Copenhagen. The environment is described as collaborative and internationally leading in NLP, with strong links to interdisciplinary social data science research. The position also includes a smaller amount of teaching and supervision. Eligibility: applicants should hold a PhD degree or equivalent in Computer Science or a related field and have strong written and oral English skills. The assessment will consider publications and relevant work experience. Experience in interdisciplinary projects, experimental research with human participants, and/or cognitive, personality, or social psychology is considered an advantage. Funding and terms: the postdoc is funded by a UCPH Convergence Grant and is offered for 1 year and 9 months . Salary starts at 39,900 DKK/month (approx. 5,300 EUR , August 2026 level), plus pension contributions and possible supplements depending on qualifications. Deadline: applications must be submitted by 5 July 2026, 23:59 GMT+1 . The expected start date is 1 October 2026 or as soon as possible thereafter. How to apply: submit the application in English through the university portal. Include a cover letter, research plan, CV, diplomas, complete publication list, three relevant papers, and contact details for three references.