Tangle & Thrive

I joined Teaching Lab Studio in 2025 as an AI Fellow to build the research practice for Tangle & Thrive, a new initiative exploring how generative AI can support student engagement and motivation. There was no research infrastructure for the project when I started, no processes, methods, or team culture around research. Building that from scratch has been one of the more interesting challenges of my career.

What we’re working on

Tangle & Thrive is focused on the following question: How do you design AI-enabled tools that students and families actually find meaningful and engaging? Our research involves interviews, surveys, and behavioral studies with students, caregivers, and educators, and feeds directly into product decisions.

One strand of work has focused on caregiver messaging: can AI-enabled SMS messages help parents support their kids’ relationship to math in ways that are low-burden but actually effective? We ran experiments with over 100 families, testing different message framings and measuring outcomes including math attitudes. Completion rates hit 80%, and we found statistically significant improvements in math anxiety and math interest.

Part of what I’ve enjoyed most is the challenge of defining what “engaging” actually means — and then figuring out how to corroborate that across different measures. It’s messier than it sounds, and getting it right matters a lot for whether the research is actually useful.

How I think about research here

Applied research in an early-stage environment is different from academic research in ways I didn’t fully appreciate until I was doing it! The timelines are shorter, the questions are messier, and the standard for “good enough to act on” is different from the standard for “good enough to publish.” I’ve found that my academic training, especially the parts about experimental design and being precise about what you’re actually measuring, transfers well. But I’ve also had to develop a different intuition for when to go deep and when to move fast.