María Grandury is a Natural Language Processing (NLP) researcher at the Polytechnical University of Madrid, dedicated to advancing responsible, inclusive, and open Artificial Intelligence, especially in the Spanish language. Her research focuses on evaluating the capabilities of large language models, with particular emphasis on social, linguistic, and cultural biases.
She enjoys sharing her research with diverse audiences and has participated as a speaker or panellist in over 35 masterclasses, talks, and round tables.
She also collaborates with organizations to increase the representation in AI for women, LGBTQ+, and LatinX communities.
She is the founder of SomosNLP, the largest Spanish-speaking NLP community, aimed at democratizing and accelerating AI advancements in Spanish by promoting the creation of open resources and accessible training. Among other initiatives, she led the development of the first leaderboard for generative large language models in Spanish and co-official languages.
María is also a Hugging Face Fellow and actively contributes to various international open-source research projects, such as BigScience and Latam-GPT, in addition to collaborating with several women-in-tech initiatives. She conducts workshops at conferences and universities and enjoys engaging in panel discussions and forums on responsible AI.
She holds a dual degree in Mathematics and Physics and is pursuing a Master’s in Language Technologies. She also has 3 years of experience at AI startups: Clibrain (Madrid), focused on Spanish LLMs, and neurocat (Berlin), focused on ML models trustworthiness assessment.
