I am a PhD candidate co-supervised by Chloé Clavel (Inria, ALMAnaCH team), Catherine Pelachaud (ISIR, Sorbonne Université), and Nicolas Rollet (Inria, ALMAnaCH team). My research investigates multimodal conversational repair strategies, specifically Other-initiated Self-repair, in human-human dialogue, examining how speakers use verbal and non-verbal cues to signal troubles and initiate repair requests.

I develop AI models for embodied conversational agents that integrate multimodal signals such as speech patterns, facial expressions, and gestures to detect human repair initiation requests and generate appropriate solutions. This work advances multimodal modeling approaches to enable more fluid conversations, minimize conversational breakdown, create more natural human-agent interactions, and increase engagement and trust in human-agent conversation.