Agentic Retrieval-Augmented Generation for Terminology Translation in Neural Machine Translation (PhD Position)
This fully-funded PhD position at South East Technological University (Carlow) focuses on advancing terminology translation in Neural Machine Translation (NMT) systems, with a particular emphasis on agentic retrieval-augmented generation (RAG) and large language models (LLMs). The project addresses the challenge of accurately translating domain-specific terms, a critical issue for translation service providers who require unambiguous and contextually consistent terminology translation in their workflows.
Despite significant progress in NMT, current models often struggle with domain-specific terminology. The research will explore innovative methods such as in-context learning with terminology examples, glossary-constrained decoding, and agentic RAG with termbases, as well as strategies for low-resource translation. The candidate will gain foundational knowledge in NLP and text analytics in the first year, then identify research objectives and develop technological innovations to advance the state of the art. Collaboration with leading research centers such as the Twin Transition Centre, ADAPT Centre, Walton Institute, and international partners (La Rochelle University, IRISA, University of Rennes) is encouraged, with opportunities for research visits abroad or in industry.
The position is funded by the SETU 2025 Presidents Scholarship Programme, offering a stipend of €22,500 per annum, full tuition fees (€5,750 per annum), and research costs (€2,000–€3,000 per annum) for four years. The ideal candidate will have a strong background in computer science, mathematics, engineering, or a related discipline, with experience in Python programming, NLP, and machine learning. Applicants must meet SETU’s English language requirements and demonstrate motivation for deep learning and computational linguistics.
To apply, download the application form from the SETU website and submit it to [email protected], quoting the reference code SETU_2025_14_CW. For informal queries, contact Dr Rejwanul Haque at [email protected]. The application deadline is April 22, 2026.
