A compelling way to close learning gaps is to support schools with a cohort of high-quality peer tutors. Schoolhouse world is already a free, online tutoring platform powered by 3k+ volunteer peer tutors that serve 25k+ learners. This project will utilize tutors in a school-centric model and leverage machine learning algorithms that improve tutor efficacy and learner assessments to double the rate of math progress for low-income middle school students.

Project Title

A Human-to-Human Peer Tutoring Platform Powered by Machine Learning

  Palo Alto, CA

Where it is used:

  • Schoolhouse is currently used in 150+ countries, primarily by students in the United States and India
About the project

What’s the problem that Schoolhouse is trying to solve?

Tutoring is one of the most effective interventions, but scaling quality tutoring is costly and challenging.

What does the intervention do?

Schoolhouse.world uses machine learning to train and match peer tutors to students, making quality tutoring more accessible to low-income students.

What is the wow factor?

Schoolhouse tutors are powered by volunteers, many of whom are high schoolers themselves, dedicated to uplifting their own peers. The strengths-based, collaborative learning model offers strong social support and training for students to learn from one another.

How does it work?

Schoolhouse will use natural language processing to suggest peer tutor moves. AI will also analyze students’ work in ways that allow downstream tasks like automatic feedback and grading. AI-enabled matching will also be used to match support tutors with learners.

or contact Drew Bent at drew@schoolhouse.world.


Chris Piech
Assistant Professor of Computer Science Education at Stanford University
Dora Demszky
Assistant Professor in Education Data Science at Stanford University
Drew Bent
Co-founder and COO of Schoolhouse.world
Matt Wu
Program & Partnership Lead at Schoolhouse.world