Postdoctoral Researcher in Trustworthy Large Language Models (m/f/d)

Main tasks

The Lamarr Fellow project of Prof. Milica Gašić, “Trustworthy Integration of Large Language Models in Human-Computer Interactive Systems", explores how we can harness the capabilities of large language models while ensuring that their use remains reliable and safe. At its core is a fundamental question: How and when can we safely integrate an LLM into a human-computer interactive system? The project investigates when LLMs should be used, when their outputs should be validated, and how uncertainty can be estimated and used to inform system decisions. An emerging direction of the project is to explore Theory of Mind and human-inspired approaches to reasoning about the beliefs, intentions, and perspectives of users and other agents, with the aim of developing more reliable and adaptive interactive systems.

Your Profile

The candidate should hold, or be close to completing, a PhD in Computer Science, Mathematics, Engineering, Computational Linguistics, or a related field. A strong background in Natural Language Processing and/or Machine Learning, together with excellent programming skills, is essential. We are looking for a curious and motivated researcher who is eager to explore fundamental questions around the trustworthy use of LLMs and to develop new ideas independently. Experience with research and publications in top-tier conferences is desirable. The candidate should be fluent in English and have excellent written and communication skills. Knowledge of German is not required.

Our Offer

The position is funded through Prof. Milica Gašić’s Lamarr Fellow project, “Trustworthy Integration of Large Language Models in Human-Computer Interactive Systems.” It is a full-time position with a competitive salary (EG 13 TV-L), offering the opportunity to contribute to cutting-edge research on trustworthy AI and human-computer interaction. The position is initially fixed-term until 31 March 2028, with the possibility of extension subject to satisfactory performance. It is a qualification position under the Act on Temporary Scientific Contracts (Wissenschaftszeitvertragsgesetz, WissZeitVG), intended to support the employee’s further scientific qualification. The start date is negotiable, with 1 January 2027 or earlier preferred.

Our Benefits

Company pension grant; Good transport connection; Free parking; Working from home; International work; Bonus payments; Union rates; Payment of capital-forming benefits

About us

The Dialog System and Machine Learning Group conducts fundamental research in Natural Language Processing and related areas of Machine Learning, with a view towards the development of the next generation of intelligent conversational agents.

Heinrich Heine University upholds the principle of "Excellence through Diversity". It has signed the "Diversity Charter" and successfully participated in the Stifterverband's audit "Shaping Diversity". It is certified as a family-friendly university and has set itself the goal of promoting diversity among its staff. Applications from people of all genders are expressly welcome. Applications from women will be given preferential consideration in accordance with the State Equal Opportunities Act. The application of severely disabled persons and persons with equivalent disabilities is also welcome. In order for a severe disability or equality to be taken into account, please provide evidence of this with suitable documents.