Inspiration
Imagine a world where education meets every child where they are, not where a standardized test says they should be. Today, children are burdened by outdated methods of assessment—methods that fail to inspire, fail to support, and fail to truly understand.
But what if education could spark joy instead of stress? What if assessment became an experience of growth and empowerment?
What it does
Joyful Minds is an AI-agent-powered platform that uses conversation to engage and assess student mastery while attending to their social, emotional, and ability needs.
How we built it
Our AI agent network uses multiple LLM agents. The UI Agent governs the conversation with the student. The Assessment Agent manages the assessment of student mastery, its confidence in that assessment, and recommendations for what to talk about with the student. The recommendation is sent to the Question Generating Agent, which crafts a question for the UI Agent and an assessment rubric for the Assessment Agent. While this is happening, the Emotional & Well-being Agent assesses the student's emotional state and engagement level and makes a recommendation to the UI Agent about how to attend to the student's needs. The UI Agent then synthesizes the question from the Question Generating Agent and the recommendation from the Emotional & Well-being Agent and responds to the student to continue the conversation. At the end of the conversation, the platform logs the mastery assessment for the student and delivers an encouraging and engaging learning plan to the student.
Challenges we ran into
Crafting the AI agent network and getting them to communicate was a challenge we overcame!
Accomplishments that we're proud of
We're proud that we engaged teachers and education administrators at our Hackathon and won huge endorsements!
What we learned
Instructing LLMs is a challenge that requires time and iterations to get right. Communication between agents is complex and requires diligence to get right!
What's next for Joyful Minds
We're revolutionizing assessment! We are field testing our assessment platform in New York City and improving our inter-agent communication
Log in or sign up for Devpost to join the conversation.