Dear all,
[Apologies for cross-posting]
The Language Technology research unit at the University of Helsinki ("Helsinki-NLP") seeks to recruit one fully funded NLP PhD student for a duration of 3 years, with extension by up to 1 extra year.
The PhD will focus on the role of pretraining data on cross-lingual transfer for factual tasks — whether data available in one language can be leveraged by multilingual LLMs to produce factually correct text in another language, and conversely, whether failure to do so can be leveraged to identify and characterize pre-training data gaps. The PhD will be conducted under the supervision of Jörg Tiedemann and Timothee Mickus; the position is expected to start on January 1st, although adjustments may be possible. The position will be held within the Helsinki-NLP team (https://blogs.helsinki.fi/language-technology/), focusing on the FLaM project (https://helsinki-nlp.github.io/flam/).
We invite applications from students currently holding a masters in computational linguistics, NLP, machine learning, or any adjacent field and not have been enrolled previously in another PhD programme. Applicants are to send an email to timothee.mickus@helsinki.fi and jorg.tiedemann@helsinki.fi with the following subject line: “[crosslingual fact transfer, PhD application]”, and must provide the following: - A curriculum vitae detailing their education (max 1 page) - The name and email of one reference (ideally a former supervisor, e.g. for a masters’ thesis) - A research statement (max 500 words), describing how the applicant’s interest aligns with the topic outlined above and what they expect to achieve through the PhD - Transcript with grades received Full consideration will be given to applications received before September 30th; the position will remain open until filled. Shortlisted applicants will be contacted directly for an interview.
We emphasize matching research interests during the assessment of applications, and will not factor in quantitative indicators (such as number of publications, etc.), in line with DORA practices.
Please forward this call to anyone who might be interested!
Best regards, Timothee Mickus Research Fellow University of Helsinki