Supporting peer interactions online and onland
UC San Diego + CMU

Current Contributors

Chinmay Kulkarni, Yasmine Kotturi, Andrew Du, Scott Klemmer.

Massive online classrooms bring together students from around the world in one connected classroom. Is it possible to teach the millions of students in an online class as well as the students in a traditional, physical classroom? Is it possible to teach them better?

We think this is possible, and that the key to this answer lies in the network of peers in these courses. Over the past three years, we have created both pedagogical styles and software systems that embed them to leverage this network of peers

Software Free and cloud-based; requires no installation

To date, tens of thousands of students have used our research systems in over 50 physical and digital classes.

  • Talkabout connects students in massive classes through scaffolded live video discussions. Talkabout discussions are incredibly diverse, most six-person discussions have students participating from five countries. Diverse discussions encourage reflection through self-reference and comparison of perspectives. To date ten massive classes and 3000+ students who have used talkabout.
  • PeerStudio provides students qualitative feedback on open-ended work. It also enables instructors to incorporate peer-supported learning with little setup. It can be used both in massive classes, and in-person ones, and has been used in multiple classes at UC San Diego.

Our vision

Google[x] invited us to give a moonshot talk; here's the 'landing'

Resources open source, and adapted widely


For other publications from our group, check out Scott's and Chinmay's research pages.


To use our software, or for press inquiries, please contact us at We also welcome your contributions to our open-source code. Here's our contributor license.

You're welcome to reuse our slides for publications in your class/reading group, and there is no need to ask for permission.

Last updated on August 30, 2016. Our research is made possible by the generous financial support of the National Science Foundation, Google, Hasso Plattner Research Program, and Media‚ÄČX.