ScaffGen

Supporting curriculum adaptions and implementation

Despite strong evidence that high-quality curriculum materials can improve student outcomes, their adoption by school districts often yields only modest impacts. A key challenge is that teachers frequently struggle to implement these materials effectively, particularly given the diverse range of student needs in today's classrooms. With approximately 50% of middle school students working below grade level, educators face significant barriers to implementing grade-level curriculum as designed.

ScaffGen is investigating how AI systems can help bridge this implementation gap. Working with expert math teachers, we have identified that educators need better tools to adapt high-quality materials for students with varying abilities, language backgrounds, and learning differences - while maintaining the curriculum's core mathematical concepts.

By analyzing patterns in how skilled teachers modify and internalize curriculum materials, we are developing targeted LLM pipelines that support common instructional workflows while preserving mathematical rigor. These insights have informed the development of Coteach.ai, an AI Assistant specifically designed for the Illustrative Mathematics curriculum. Coteach utilizes a multi-agent system with specialized agents for curriculum analysis, content creation, and diagram visualization. Currently serving approximately 1,000 educators across three states, the ScaffGen project aims to advance our understanding of how AI tools can support curriculum coherence and educational equity while respecting teacher expertise and judgment.

Visit the Coteach.ai Homepage

Awards 

Winner Accelerating and Assessing Learning Track, 2024-2025

Seed Grant Recipient, 2024