Jinsook (Jennie) Lee

I’m a Ph.D. candidate in Information Science at Cornell University advised by René F. Kizilcec in the Future of Learning Lab and Thorsten Joachims. My research studies socio-technical systems in education through the lens of computational social science, focusing on how AI technologies influence evaluation, decision-making, and equity in high-stakes contexts.

At Cornell, I study how language models influence college admissions both directly through applicants’ use of LLMs in essay writing, and indirectly, through how institutions interpret and act on these AI-mediated signals. My recent work analyzes (1) how policy shifts and the rise of LLM-assisted writing reconfigure the allocation of educational opportunities; (2) how LLMs write differently in terms of lexical diversity, semantic space and stylistic homogenization in college essays; and (3) how uncertainty and arbitrariness manifest in algorithmic predictions within admissions.

Beyond admissions, I am also developing AI evaluation pipelines for tutoring data in collaboration with National Tutoring Observatory spanning multi-agent orchestration, and dialogue segmentation to enhance tutoring move annotation.

I’m also fortunate to collaborate with Nikhil Garg and AJ Alvero.

Prior to Cornell, I spent several years as a data scientist at Korea University to develop course and major recommender systems to support college students’ decision making process.

I have a love-hate relationship with tennis — You’ll often find me attempting to upgrade my skills from the ‘absolute beginner’ category. I also love listening to music and curating songs!