Mary Osborne, Lecturing Fellow

Mary Osborne is a Senior Product Manager for Natural Language Processing and Generative AI at SAS and a Lecturing Fellow in Computational Linguistics at Duke University. Her work bridges industry innovation and academic inquiry, exploring how linguistic theory, cognitive science, and machine learning converge in the design of trustworthy, interpretable AI systems.

At Duke, Osborne teaches courses such as Computational Approaches to Human Language and Artificial Intelligence, Linguistic Theory, and Large Language Models, guiding students through the historical and conceptual foundations of language technologies—from Saussure and Shannon to transformers and large language models. She also leads the Focus Program cluster Implications of AI and Large Language Models, an interdisciplinary initiative connecting linguistics, philosophy, computer science, and education to examine how artificial intelligence reshapes our understanding of mind, meaning, communication and the landscape of education.

At SAS, Osborne drives product strategy for NLP and GenAI capabilities across the SAS Viya platform, focusing on responsible AI, linguistic model evaluation, and enterprise applications of text analytics. Her professional and academic work share a common goal: to make AI systems not only powerful, but explainable, ethical, and aligned with human understanding of language.

Contact Info: 

Office Location:
Office Phone:	(919) 660-3140
Email Address:
Web Page:	https://blogs.sas.com/content/author/maryosborne/

Teaching (Spring 2026):

• LINGUIST 399.01, COMP APP TO HUMAN LANGUAGE Synopsis

  Old Chem 116, MW 06:15 PM-07:30 PM

  (also cross-listed as COMPSCI 376.01)

Education:

B.S.

University of North Carolina, Wilmington

1998

Keywords:

Artificial Intelligence • Computational linguistics • Deep learning (Machine learning) • Machine learning • Natural Language Processing • Natural language processing (Computer science) • Predictive analytics • Text data mining • Transfer learning (Machine learning)