معرفی
Dorottya Demszky is an Assistant Professor in Education Data Science at Stanford University's Graduate School of Education, with a courtesy appointment in the Computer Science Department. She leads the EduNLP Lab, an interdisciplinary research group focused on developing natural language processing methods to support equitable and student-centered instruction. Her work bridges education, computer science, and linguistics to create interpretable and scalable educational measures.
Dr. Demszky earned her PhD in Linguistics from Stanford University under Dan Jurafsky's supervision and completed her BA summa cum laude in Linguistics with a minor in Computer Science at Princeton University. Her educational background uniquely positions her at the intersection of language, technology, and education.
Her research focuses on developing NLP tools for educational contexts, including systems that provide feedback on dialogic instructional practices, analyze representation in textbooks, and measure dialect features in student writing. She combines machine learning, natural language processing, and linguistics with practitioner input to develop tools that improve educational interventions, teacher professional development, and curriculum equity. Her recent work shows increasing sophistication in adapting NLP for noisy classroom environments and supporting multilingual learners.
Dr. Demszky's publications demonstrate a clear trajectory from foundational NLP methods for education toward increasingly practical classroom applications. Her 2023-2025 work shows growing emphasis on real-world implementation, with multiple randomized controlled trials validating the impact of her tools on teaching practices and student outcomes. The research spans theoretical advances in NLP for education to practical tools deployed in classrooms.
- Best Dataset Prize at Learning at Scale (MathemaTikZ)
- IEDMS Publicly Available Educational Dataset Prize (NCTE classroom transcript dataset)
- Selected as a Leading Woman in AI at ASU GSV AIR Show
- NSF RAPID grant for understanding how instructional coaches integrate automated feedback
- $70k seed grant from Stanford HAI for adapting mathematics curricula
- $100k Gates Foundation grant for equitable speech recognition systems
Dr. Demszky actively advises PhD students in the Learning Sciences and Technology Design program and has secured significant funding from the NSF, Gates Foundation, and Stanford HAI. Her lab's Practitioner Voices Summit brought together 60 teachers from 22 states to inform AI research for classrooms. She is particularly committed to ensuring educational technologies are developed with and for teachers, emphasizing social responsibility in edtech design. Her work with the Tarisznya Alapítvány (Knapsack Foundation) in Hungary demonstrates her long-standing commitment to educational equity for underprivileged children.
The EduNLP Lab develops tools that gather evidence about effective teaching practices, develop validated algorithms through co-design with educators, and pilot solutions that teachers can immediately use. Current projects include CoTeach.AI, which helps teachers scaffold math curricula, and systems for analyzing classroom dialogue to improve teacher-student interactions. The lab emphasizes human connections first, using technology to enhance rather than replace the essential relationships in education.
