PhD Student – Mechanistic Interpretability
Job Reference: 743_25_LS_LT_R1
Closing Date: Thursday, 18 December, 2025
About BSC
The Barcelona Supercomputing Center – Centro Nacional de Supercomputación (BSC‑CNS) is the leading supercomputing centre in Spain. It hosts MareNostrum, one of Europe’s most powerful supercomputers, and serves as a founding member of the former European HPC infrastructure PRACE as well as a hosting entity for EuroHPC JU. BSC’s mission is to research, develop and manage information technologies to facilitate scientific progress. The centre combines HPC services and R&D in computer and computational science across life, earth and engineering sciences under one roof, with over 1000 staff from 60 countries.
Context And Mission
“Multilingual Production Principles of Lexical Collocations” (MaPPLexiC) is a five‑year project funded by the European Research Council (2025–2030). Leveraging recent advances in computational lexicography, cognitive science and neural computational linguistics, MaPPLexiC seeks to uncover and articulate in human‑readable rules why specific word combinations create collocations with particular meanings, by interpreting neural collocation recognition and classification models.
Key Duties
* Explore existing methods for mechanistic interpretability and research (design and develop) novel techniques for this task.
* Contribute to the design and implementation of feature‑visualization techniques in neural representations.
* Participate in cross‑language evaluation studies.
* Contribute to the improvement of the performance of existing neural collocation recognition and classification models.
* Collaborate with the corpus annotation team.
Requirements
* Education
o Master in Computational Linguistics, Computer Science, Machine Learning, Mathematics, or Physics.
* Essential Knowledge and Professional Experience
o Solid theoretical knowledge in neural language modelling and natural language processing.
o Strong programming skills in Python and deep‑learning frameworks.
o Ability to work effectively in a collaborative, multidisciplinary environment.
* Additional Knowledge and Professional Experience
o Expertise in implementing deep machine‑learning‑based applications.
* Competences
o Fluency in spoken and written English.
o Competence in one or several project languages (Czech, Finnish, French, German, Hungarian, Russian, Spanish). Any additional language is an advantage.
Conditions
* The position will be located at BSC within the Life Sciences Department.
* We offer a full‑time contract (37.5 h/week), a stimulating environment with state‑of‑the‑art infrastructure, flexible working hours, extensive training plan, restaurant tickets, private health insurance, and support for relocation procedures.
* Duration: 1 year.
* Holidays: 22 days + 6 personal days + 24 th and 31 st December per the collective agreement.
* Salary: Competitive, commensurate with qualifications and experience and according to the cost of living in Barcelona.
* Starting date: 01/01/2026.
Applications Procedure and Process
* Submit through the BSC website.
* A full CV in English, including contact details.
* A cover/motivation letter in English stating the specific area and topics of interest, and including two reference contacts.
* Applications lacking any of these documents will not be considered.
Recruitment Process Development
* Curriculum Analysis: Evaluation of experience, degree, training and other relevant information (40 points).
* Interview phase: Assessment of technical competencies, knowledge, skills, personal competencies, and a minimum of 30 points out of 60 required for eligibility (60 points).
* The recruitment panel will comprise at least three people, with a minimum 25 % representation of women.
* All candidates invited to interviews will receive feedback on the outcome.
BSC‑CNS is an equal‑opportunity employer committed to equity, diversity and inclusion. We consider all qualified applicants for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability or any other legally protected characteristic.
#J-18808-Ljbffr