Elyas Meguellati

I'm a researcher working at the intersection of generative AI, persuasion, and natural language processing. I study how language models shape human decisions and how we can design them to be more effective, efficient, and aligned with real-world needs.

Right now, I'm completing my PhD at The University of Queensland, where my work focuses on the duality of persuasion — how AI models generate tailored messages and how we can detect persuasive techniques in the wild. Along the way, I've engineered lightweight classifiers, multimodal pipelines, and LLM-driven systems that personalize content at scale.

I've led collaborations with industry partners such as Allianz Worldwide Partners, building personalization engines and experimenting with generative models in real deployments. I also work within the ARC Centre of Information Resilience, developing context-aware models for harmful content detection and semantic augmentation.

My broader goal is simple: build AI systems that are powerful, efficient, and grounded in human behaviour. Systems that don't just optimize accuracy, but understand nuance, context, and impact.

If you're interested in generative AI, persuasion, or applied NLP — or if you just want to talk ideas — I'm always happy to connect.