Offer Description
UCLouvain is looking for a doctoral researcher (F/M/X) for a project that involves natural language processing (NLP) and automatic term extraction (ATE).
- Full-time doctoral fellowship (100%) for two years, with conditional renewal for another two years
- Institute for Language & Communication (ILC) at UCLouvain (Louvain-la-Neuve, Belgium)
- Start date: after 1st October 2025 (the start date is flexible)
Short project description:
Specialised bilingual glossaries are vital resources for the translation industry. Their development is, however, very time-consuming, and bilingual parallel corpora that can serve as basis for their creation are scarce, depending on the language pair and the specialised domain. This project thus focuses on developing a new methodology, based on existing machine learning and deep learning techniques, for automatically extracting bilingual terminology using comparable corpora. By combining language models fine-tuned for token classification and cross-lingual sentence similarity, this project proposes a cost-effective way of automatically building specialised bilingual glossaries on demand.
To achieve that, the PhD project consists of two main goals:
1. Create a bilingual comparable corpus in French and English, focusing on the general domain of international patents, which has to be semi-automatically annotated with terms and bilingual terminology alignments. The specific domain of the corpus will be defined by the doctoral researcher.
2. Create a new methodology for automatic terminology extraction (ATE) using deep learning and fine-tuning of existing language models. This objective is divided in three main parts:
a. Testing the existing methodologies for ATE;
b. Analyzing positive and negative aspects of the existing methodologies; and
c. Developing a new methodology for ATE that can be adapted to other languages.
The doctoral researcher is expected to conduct the essential points of the research to attain the above-mentioned goals. They should also ensure the visibility of the results by publishing the work in academic venues and write a doctoral thesis.
Work environment
The Institute for Language & Communication (https://www.uclouvain.be/en/research-institutes/ilc) is part of UCLouvain (https://www.uclouvain.be/en/the-university), a lively university with its main campus in Louvain-la-Neuve, a city with a very dynamic, pedestrian-only urban centre that offers good quality of life and is very close to Brussels (around 30km).
The research project will be conducted under the supervision of Leonardo Zilio, whose specialty lies in the study of specialized language and terminology extraction.
Where to apply
E-mail [email protected]
Website https://jobs.uclouvain.be/Personnelscientifique/?locale=en_GB
Requirements
Research Field Language sciences
Education Level Master Degree or equivalent
Skills/Qualifications
The applicant should have the following qualifications and skills:
- A master’s degree in computational linguistics, natural language processing, computer science (with emphasis in artificial intelligence), digital humanities, or a related domain.
- An excellent academic curriculum.
- Good knowledge of French (at least B2 level) and English (at least B2 level)
- Good computational and programming skills:
* Programming language: Python
* Knowledge of scikit-learn, pandas, tensorflow/keras and/or pytorch.
- Good knowledge of the main NLP tools. Knowledge of deep neural networks and token-classification models is a plus.
- Knowledge of lexical sciences, specially terminology.
- Autonomy, curiosity, collaborative sense, good attention and analytical skills, proactivity and reactivity.
Application documents
Application deadline: 17 August 2025
If you are interested in this doctoral fellowship, please send your application per e-mail to Leonardo Zilio ([email protected]) and attach all the requested documents below. These are the requested documents:
1. A detailed curriculum vitae in English, showing the required qualifications and skills, the academic record (with courses and final marks), as well as existing publications (if any), work experience (if any) and other academic and scientific experience;
2. A motivation letter, in French, describing why the applicant is interested on this doctoral fellowship and how their profile aligns with the project description and its objectives (max. 2 pages);
3. A reference letter in English or French from a previous lecturer / supervisor.
4. An concise academic statement in English, indicating the expectations of the applicant in relation to their doctoral studies and their career objectives.
The pre-selected applicants will be invited to attend an online interview. Details for the interview will be sent to those applicants per e-mail.
More information:
Questions regarding the fellowship and the application procedure can be sent per e-mail to Leonardo Zilio ([email protected]) before the 25th of July 2025.
Specific Requirements
Engagement conditions
The offer of this doctoral fellowship is subject to the following conditions:
* After the recruitment, the selected applicant should also be approved by the UCLouvain’s Research Counsel
* The doctoral researcher should live in Belgium for the whole duration of the fellowship.
* The net monthly amount that can be expected by the doctoral researcher is around 2,450 euros.
* The initial contract is for two years. To ensure the renewal of the fellowship, the doctoral research will have to successfully pass the mid-term evaluation of the doctoral thesis and also apply for FNRS funding and attain a classification at rank “A”.
* The maximum period indicated above is increased by one year in case of paternal or maternal leave.
Languages ENGLISH
Level Good
Languages FRENCH
Level Good
Research Field Language sciences
Additional Information
Work Location(s)
Number of offers available 1
Company/Institute Institute for Language and Communication
Country Belgium
City Louvain-la-Neuve
Postal Code 1348
Street Place cardinal Mercier 31
Contact
City Louvain-la-Neuve
Website http://www.uclouvain.be/ilc
Street Place Cardinal Mercier 31
E-Mail [email protected]
