AI Model Solution for Tutoring

The agency is seeking to fund the creation of open-source, K-12 education-specific Artificial Intelligence model solutions and supporting research to enable AI math tutoring that is as effective as human experts. This initiative aims to improve the use of technological innovations in tutoring environments. The project covers deploying the model via a standard AI inference stack such as vLLM or SGLang, integrating with an ed-tech partner for product-level evaluation and deployment, documenting implementation playbooks, conducting risk evaluations, and performing cost analyses. Only organizations based onshore in the United States are eligible to apply for this opportunity. The agency requires all performance of the work to be conducted offsite. Bidders must submit their responses digitally either via email or through the online portal. The agency has established a budget of up to 8,000,000 USD for this grant. The source document does not specify evaluation criteria or key dates for this opportunity.