I am an associate professor (reader according to the UK system) at Heriot-Watt University. I am also a co-leader of the National Robotarium Safe and Secure AI for Robotics (SAIR) theme.

My main research goal is to build Responsible, Safe, and Explainable AI Applications using Natural Language Processing (NLP).

My current interests revolve around the detection and mitigation (through Counter-speech) of Gender Based Violence (GBV) using LLMs, and Social Behaviour Modelling in LLMs. More recently I have been looking into interpreting the reasoning capabilities and knowledge capacity of LLMs mechanistically, and I have been exploring the linguistic role of instructions in Vision-Language-Action (VLA) Models for robotic manipulation.

Of course, my long-standing interest has always been in Natural Language Generation (NLG) in terms of representations, modelling and evaluation. Throughout the years I have investigated generating text from meaning representations, programming language code, generating stories from sequences of generic events, or other educational domains such as Math word problems, as well as Dialogue Systems and Conversational Agents embedded in autonomous robots.

I continuously enjoy investigating new research fields, and hence, I have also acquired skills in Psycholinguistically Motivated Parsing, semantics, and Information Retrieval from my past employment.
I was a postdoctoral researcher in the Department of Computer Science & Engineering at the University of Washington, working with Luke Zettlemoyer and Yejin Choi, between 2015 and 2017.
I have also worked as a research associate (post-doc) at the University of Edinburgh, with Frank Keller and Mirella Lapata.
I completed my PhD in 2013, 'Joint Models for Concept-to-text Generation', at the University of Edinburgh, supervised by Mirella Lapata.