Farid Adilazuarda

I’m a PhD student at the University of Edinburgh, where I’m fortunate to be advised by Edoardo Ponti and Alexandra Birch. I’m a member of the Institute for Language, Cognition and Computation (ILCC) and the Edinburgh NLP Group.

My research asks a big question: “How can we design language models that are both compute-efficient and multilingual, so they can reason better under tight budgets and across languages?” To answer this, I focus on improving the efficiency of neural models, with a particular interest in methods that could enable tokenization-free and more compute-optimal architectures. Currently, I work on adaptive memory compression for faster inference and multilinguality, aiming to build models that perform well across diverse languages.

Before Edinburgh, I was a Research Assistant at MBZUAI in Abu Dhabi and a visiting researcher at the UKP Lab, TU Darmstadt, where I worked on multilingual, multicultural, and parameter-efficient NLP—topics I still enjoy discussing!

I know how much research success can depend on access and opportunity. If you’re from an underrepresented group and think a conversation might help, feel free to reach out by email.