Postdoctoral Researcher in Explainable Multimodal AI

About the role

As a Postdoctoral Researcher at TU Delft you will explore how multimodal AI systems can understand, explain and communicate predictions about human behaviour. The work focuses on interactive, real‑time explanation in situated conversational settings, with applications in augmented and mixed reality environments.

Key responsibilities

• Investigate Vision‑Language Models (VLMs) and Multimodal Large Language Models (MLLMs) for explainable AI.
• Develop mechanistic interpretability methods and concept‑based explanations grounded in affect and behaviour theory.
• Design and implement multimodal sensing pipelines using vision, audio and physiological signals.
• Create interactive systems that support real‑time dialogue and user questioning in immersive environments.
• Collaborate with the Interactive Intelligence Group and the Hybrid Intelligence Centre, attending consortium meetings and partner visits.

Required profile

• Hold, or be expected to obtain before the start of the project, a PhD degree in a relevant discipline.
• Demonstrated research track record in AI, machine learning, or related fields.
• Strong interest in human‑AI interaction and explainable AI.

Required skills

• Experience with Vision‑Language Models (VLMs).
• Experience with Multimodal Large Language Models (MLLMs).
• Knowledge of multimodal sensing techniques (vision, audio, physiological signals).

What we offer

• Inclusive research environment with tailored mentorship and training.
• Opportunities to work within a collaborative group and a Dutch consortium on Hybrid Intelligence.
• Access to state‑of‑the‑art facilities for immersive and mixed‑reality research.