Katarzyna Foremniak is a computational linguist with over 10 years of experience in NLP and speech recognition. She has developed language models for automotive brands like Audi and Porsche, working on the voice interfaces and language technology that drivers interact with every day, and her expertise bridges linguistics and machine learning in equal measure.

Her specialization covers phonetics, morpho-syntax, and sentiment analysis, and in recent years she has focused on human-centered speech recognition: making automatic speech recognition work for atypical and accented speech. That includes strategies for recognizing disordered speech, such as specialized datasets, data augmentation, synthetic variations, and transfer learning with limited data, as well as the data collection challenges that come with GDPR and clinical data.

Katarzyna also teaches at the University of Warsaw, passing on her knowledge of linguistics and language technology to the next generation. She is passionate about human-centered AI and multilingual NLP, and she discussed assistive speech recognition applications and human-centered AI on the DataTalksClub podcast.