Postdoc in AI-Driven Adaptive Learning Systems (CLARA project) Eindhoven University of Technology in Netherlands

Eindhoven University of Technology (TU/e) is recruiting a postdoctoral researcher for the NRO-funded CLARA project, AI as a Social Agent to Support Group Learning Processes. The position is hosted in the Department of Industrial Engineering & Innovation Sciences (IE&IS) and focuses on designing an LLM-based social agent that can support students’ collaborative learning in Challenge-Based Learning environments.

The project sits at the intersection of machine learning, natural language processing, human-computer interaction, and educational research. As the technical core of the CLARA consortium, you will work closely with a PhD candidate and an interdisciplinary supervisory team, while collaborating with partners at the University of Twente and Maastricht University. The role involves designing, training, and iteratively refining an AI agent that can detect socio-cognitive and socio-emotional triggers in group work and deliver contextually appropriate scaffolding in real time.

Research tasks include building trigger detection systems for text, audio, and multimodal interaction data; fine-tuning large language models on educational datasets; designing adaptive scaffolding interventions; supporting pilot studies in real CBL classrooms; and conducting responsible AI work such as fairness validation, data governance documentation, and GDPR compliance. The postdoc will also contribute to open science outputs, containerized model workflows, publications, conference presentations, and knowledge transfer activities within the consortium.

The post is suited to a technically strong researcher with a PhD in Computer Science, Artificial Intelligence, or a closely related field, plus demonstrable experience with LLMs, NLP, machine learning, Python, and modern ML tooling such as PyTorch, HuggingFace Transformers, and LangChain. Experience with fine-tuning, prompt engineering, retrieval-augmented generation, agentic AI, and HPC environments is explicitly requested. Strong English communication skills and the ability to collaborate across engineering, innovation sciences, and educational research are important. Experience with educational data, multimodal signals, and active learning environments is an advantage.

Employment is for 2.5 years at 0.8 FTE (32 hours/week), with salary in scale 10 of the Dutch Universities collective agreement: €4,422 to €5,760 gross per month, plus an 8.3% year-end bonus and 8% vacation pay. TU/e also offers training programs, supervision and mentoring, an international research environment, technical infrastructure, support for international staff, and additional benefits such as commuting and home-working allowances.

The application deadline is 17 August 2026. Applicants must submit a cover letter, CV including two or three references, one representative publication or code portfolio, and a short technical challenge response. The challenge asks candidates to identify a trigger event from speech-based student-state data and generate a suitable scaffolding response, together with a brief reflection on educational appropriateness and ethical limitations.