Postdoc position Multimodal Conversational AI Cultural Heritage

Looking for a job that matters? Join the university of technology that puts people first – and shape new opportunities both for yourself and for ou...

We are looking for a postdoctoral researcher to work on the project Balancing Personalisation and Polyvocality:
Dialogic Interaction for Open-mindedness and Empathy in Conversational AI. The project investigates how conversational agents in museum settings can balance personalisation with polyvocality: adapting to individual visitors while also encouraging engagement with unfamiliar, marginalised, or challenging perspectives.

Many current conversational agents and recommender systems adapt content to users’ existing interests and preferences. While this can improve engagement, it may also narrow what users encounter. In museum contexts, this is a particularly important challenge. Museums aim not only to inform and engage visitors, but also to support reflection, openness, empathy, and cross-cultural understanding.

In this project, you will develop and evaluate a multimodal, polyvocal conversational agent. The agent will use real-time visitor feedback, including eye gaze, speech, and dialogue content, to infer interactional states such as engagement, curiosity, openness, confusion, or resistance. These signals will be used to design adaptive dialogue strategies that maintain engagement while introducing diverse perspectives in a meaningful and responsible way.

The position will be embedded in the Human Media Interaction group at the University of Twente and will contribute to the Hybrid Intelligence Centre . You will work in an interdisciplinary research environment at the intersection of conversational AI, human–computer interaction, multimodal interaction, user modelling, and cultural heritage.

As a postdoctoral researcher, you will:

• Conduct research on dialogic interaction, multimodal conversational AI, user modelling, polyvocality, and cultural heritage interpretation;
• Develop and evaluate multimodal models that use gaze, speech, and dialogue data to infer visitor states such as openness, curiosity, engagement, confusion, and resistance;
• Design and compare adaptive dialogue strategies that balance personalization with exposure to diverse perspectives;
• Conduct VR-based user studies to evaluate the effects of adaptive dialogue strategies on engagement, openness, curiosity, empathy, and reflection;
• Contribute to reusable experimental infrastructure, datasets, evaluation methods, and software components;
• Publish research results in high-quality international journals and conferences;
• Collaborate with researchers, a PhD candidate working on polyvocal knowledge representation, curators, museum professionals, and other partners in the Hybrid Intelligence network.

• A PhD in computer science, artificial intelligence, human–computer interaction, computational social science, digital humanities, cognitive science, or a related field;
• Experience with conversational AI, dialogue systems, human–agent interaction, adaptive systems, user modelling, multimodal interaction, or a related area;
• Experience with empirical research, preferably including user studies;
• Strong analytical and academic writing skills;
• The ability to work independently and as part of an interdisciplinary team;
• Excellent communication skills in English.
• Experience with one or more of the following is an advantage:
• Multimodal interaction analysis, eye tracking, speech analysis, or affective computing;
• VR-based experiments or immersive interaction;
• Adaptive dialogue management, reinforcement learning, or decision-making for conversational agents;
• User modelling or computational modelling of engagement, curiosity, openness, or resistance;
• Knowledge graphs, ontologies, semantic technologies, or Linked Data;
• Cultural heritage, museum studies, digital humanities, or participatory design;
• Collaboration with non-technical domain experts such as curators or museum professionals.

We especially welcome candidates who can connect technical development with empirical evaluation of human–AI interaction, and who are enthusiastic about working…