I am a Research Investigator and Lecturer IV at the University of Michigan School of Information, where I teach data science to undergraduate and graduate students and design the assessments and tools those courses run on.
My teaching and research explore how AI can make assessment continuous, embedded in the learning process, and useful to learners, rather than a high-stakes event at the end of it.
Most assessment is still summative, measuring students under pressure, at a single moment, against a single standard. That approach often misreads what learners know, offers little feedback they can act on, and tends to widen existing inequities. AI has not yet solved this, and hyper-personalized tutors introduce a problem of their own: although they promise to meet each student where they are, they turn learning into a private exchange between one student and a machine, leaving out the collaborative, social processes through which people build knowledge together.
I design systems that work differently, tuning assessment to what students are trying to understand and achieve, including goals that traditional tests never capture. In my own data science courses, for example, students demonstrate understanding by explaining and defending their analytical reasoning, rather than only by submitting code that runs.
My work draws on the learning sciences, data science, and social learning, with an emphasis on design-based implementation, in which I build educational technologies, deploy them in real classrooms, and refine them based on what happens there.
Current teaching
- SI 345: Data Manipulation with Python
- SI 385: Exploratory Data Analysis and Visualization
- SI 618: Data Manipulation and Analysis
Elsewhere on this site
The projects page describes the assessment systems and tools I have built, writing collects longer pieces on teaching data science now that AI can write the code, and work with me lists the kinds of consulting and workshops I take on.