AI Developer Tools for Education

We’re working to realize AI’s full potential in education by building foundational AI infrastructure to accelerate the development of more accurate, state academic standards-aligned, and learning-science backed edtech products for the public good.

Our goal is to address the gap between big tech's AI advancements and the specific needs of education. With our developer tools, we aim to empower edtech teams to create transformative and pedagogically rigorous solutions that address the unique challenges faced by educators and schools

Partner With Us

Our Approach

We’re co-building with partners to improve AI for education.

We believe that by building AI systems with researchers, educators, and other leaders in the space, we can help drive a future that is focused on better serving students and teachers and one that is accessible and engaging for all. 

For our initial private betas, we're partnering with select edtech developers to test and iterate on tools and resources that will help evaluate and improve the quality of their AI outputs and products.

Become a Partner

Initial Focus Areas

Our first two tools help developers evaluate and improve the performance of the educational AI outputs used in their edtech products.

Knowledge Graph Evaluators

Our Knowledge Graph helps developers improve the quality of outputs from educational AI systems.

We’re working to increase the impact of AI for education by improving the effectiveness and accuracy of AI systems through high-quality inputs.

Our Knowledge Graph delivers a structured network of datasets across curricula, state academic standards, and learning science research, enabling developers to improve the accuracy of their AI outputs.

Selecting the Right Source of Education Data Providers

To build out a knowledge graph that maps curriculum and standards, we start by establishing partnerships with state academic standards and high-quality curricula providers.

Constructing a Graph From High-Quality Data

We then convert the data into a consistent, semantically rich format that is relevant and reliable.

Enabling the Graph for Developers To Utilize

Correct information can then be pulled through a variety of techniques directly from the knowledge graph into a developer’s LLM context window.

Our Initial Release:
Illustrative Math Knowledge Graph

Our aim is to provide developers with datasets that support scalability and impact by leveraging high-quality content. For our initial knowledge graph, we’re incorporating state academic standards and proven curricula. Our team is starting with the Illustrative Mathematics curriculum, which is widely used in K–12 education, due to its strong foundation in learning science, focus on rigor, and alignment with educational standards. We are also incorporating academic standards from all 50 states in partnership with 1EdTech.

We’re partnering with Playlab in a closed beta, where they will leverage our knowledge graph to increase the effectiveness of their platform across the entire learning experience using Illustrative Math.

Learn About Our Evaluators

Our Evaluators enable developers to assess the quality of educational AI outputs.

We’re working to measure GenAI’s quality in order to make outputs more pedagogically aligned.

Our Evaluators enable developers to ensure high-quality AI-generated content by assessing outputs against expert-backed rubrics. These scalable, cost-effective solutions allow teams to prioritize building dependable products and confidently ship their work.

Choosing Comprehensive
Open-Source Rubrics

We select expert-backed rubrics that capture key dimensions of pedagogy.

Creating Validation 
Datasets

We then collect product outputs as datasets from other GenAI tools to be annotated by domain experts against the selected rubric.

Configuring the
Optimal LLM

After annotation, we optimize the dataset across various LLMs to identify the optimal model configuration for scoring inputs against the rubric, achieving accuracy comparable to human experts.

Our Initial Release:
Qualitative Text Complexity Evaluator

As a starting point, we are focusing on reading text complexity to help close the gap in student reading skills. Edtech developers need reliable tools to ensure the texts they deliver meet teachers' desired complexity and rigor. Our evaluator leverages the SCASS rubric from Student Achievement Partners, and we collaborated with a team of English Language Arts experts from The Achievement Network and Gradient Learning for dataset assessment. These partnerships ensure our evaluator can provide product teams with a dependable evaluation of text complexity.

We’re partnering with Diffit in a private beta, in which they will use our evaluator to measure and improve the complexity of the text their product generates.

Learn About Our Knowledge Graph

Privacy by
Design

Our tools are built with privacy as a core principle, along with strong security measures. By prioritizing privacy, we handle data responsibly, maintaining trust while delivering valuable educational resources. We also provide clear and comprehensive explanations about how our tools operate, ensuring educators, developers and partners can confidently integrate them into their work.
Development of
AI-powered Tools

We work collaboratively with educators, researchers and other domain experts to design AI tools that are inclusive and aligned with the real-world needs of classrooms. We believe AI is a tool meant to complement human expertise, not replace it.
Our AI
Governance Program

As a part of AI Developer Tools for Education, we build resources guided by our AI governance principles, including protecting confidentiality, privacy and security; upholding accountability; and providing transparency. Our AI Governance program includes processes about data usage, AI development, and AI use. Through our governance principles, we maintain a transparent and ethical approach to building AI-powered educational tools.