Postdoctoral Position: Latent Variable Models for Creative Language Generation KU Leuven in United Kingdom

KU Leuven is recruiting a postdoctoral researcher for the ERC Consolidator Grant project TENACITY (Tensors and Neural Networks for Computational Creativity) at the Centre for Computational Linguistics (CCL), part of the ComForT research unit. The project explores computational models of language that exhibit creativity, with a specific focus on integrating latent factorization models and neural language models for creative language generation.

The research agenda brings together two complementary lines of work: interpretable tensor factorization models that induce latent semantic structure from multi-way word co-occurrences, and constraint-based neural language models designed to steer generation toward creative output. The successful candidate will help build an integrated framework in which latent structure from factorization models informs the latent representations of transformer-based language models, potentially through variational inference approaches. The goal is to better control and understand latent variables so they can be used for creative generation, including tasks such as producing text in a particular style or other expressive forms of language use.

This is a challenging high-risk/high-gain project with substantial freedom for the postdoctoral researcher to shape the research direction according to their expertise and vision. The position is research-only and has no teaching obligations.

Responsibilities include developing variational inference methods for transformer-based language models informed by tensor factorization structure; investigating interpretability of the latent space and its use for creative language generation such as metaphor and style; collaborating closely with the project’s PhD students and providing day-to-day guidance; and publishing results in top NLP and machine learning venues and presenting at international conferences.

Profile: applicants should hold a PhD in natural language processing, machine learning, computational linguistics, or a related field, ideally with peer-reviewed publications. Strong deep learning expertise in NLP is expected, along with a strong plus for experience in latent variable models and/or variational inference. The role also requires excellent Python and PyTorch programming skills, experience with large-scale experimentation on GPU infrastructure, and strong English communication skills. Experience guiding junior researchers is welcome.

Offer and funding: the position is full-time for 2 years in a high-level ERC-funded research environment, with access to ample computing infrastructure and funding for conference travel. KU Leuven emphasizes an inclusive and respectful environment and encourages applications from diverse candidates.

Application and deadline: apply via the KU Leuven jobsite link. The application deadline is 2026-08-10 23:59 (Europe/Brussels).