ChatEDU is an edtech startup leveraging generative AI to create personalized learning tools for college students. Over the last nine months, we've developed three products that have won five hackathons, earned funding and mentorship from Microsoft, and attracted a team of senior advisors.

• ChatEDU V1 offers a gamified, socialized AI tutor for college students
• ChatEDU V2 extends the assessment framework from v1 to offer an intelligent learning management system with comprehension analytics for teachers
• ChatEDU V3 combines the capabilities of the first two products to offer a dynamic copilot for students that learns and works with them, not for them

Through our collaboration with Vanderbilt’s Initiative on the Future of Learning & Generative AI, we are amongst the first students to apply knowledge graphs, semantic search, and multi-agent systems to the field of education.

We aim to create the first viral higher education product since Quizlet and Chegg. As college students, we witnessed the transformative impact of ChatGPT on our peers and were inspired to create deeply personalized educational tools. Our multidisciplinary engineering backgrounds and direct insights into student pain points equip us to tackle the $5 billion edtech market.

Jason Hedman CS + Math

JP Higgins CS + HOD

Vasco Singh CS + Math

Jake Underwood Engineering Management + CS

During our last year at Vanderbilt, we built three edtech products, incorporated feedback from 50 student users across six universities, and won five hackathons. Powered by this success, we are now pursuing ChatEDU full-time.

Gamified, socialized AI tutors grounded in each student’s learning materials.

🥇 Winner of Microsoft AI Classroom Hackathon, Vandyhacks X, Cornucodia

https://github.com/chat-edu/chat-edu

Technology: RAG-powered question generation (Azure OpenAI, Azure Cosmos DB, Azure AI Search), topic extraction + note modularization (Azure AI Language)

Student Feedback:

• Strengths: Our platform's gamified social elements made the mundane studying process fun and engaging. Sharable notebooks drove growth and promoted the creation of high-quality learning materials.
• Weaknesses: The market for study tools is saturated, and many other incumbents and startups are integrating similar functionalities. Generative AI creates a new range of educational possibilities beyond automation.

Intelligent learning management system offering comprehension analytics, lesson plan and assignment generation, and actionable feedback for classrooms.

🥇 Winner of Vanderbilt Generative AI Showcase

https://github.com/chat-edu/chat-edu-v2

Technology: comprehension agents, knowledge graphs, adaptive assignment creation

Teacher Feedback:

• Strengths: Teachers spent less time on paperwork and more time addressing their students' unique needs. Knowledge graphs on the individual and class-wide levels allowed them to identify and address learning gaps.
• Weaknesses: A full-scale school deployment would require navigating long sales processes and meeting intensive compliance requirements. Incumbent learning management systems hold long contracts with schools and offer similar capabilities.

Multimodal copilot that guides you through complex learning objectives.

🥇 Winner of Microsoft Generative AI Hackathon

https://github.com/chat-edu/second-brain

Technology: Multimodal note, lecture, and assignment processing (Azure AI Video Indexer, Azure AI Vision, Azure AI Speech, Azure AI Document Intelligence, GPT-4o), knowledge graph construction and validation agents (Azure OpenAI Assistants), dynamic comprehension profiles

Student Feedback:

• Strengths: Knowledge graphs structure otherwise scattered materials, illustrating the interconnectedness of course topics. Storing progress in these graphs provides further personalization and encourages further use. Breaking tasks down into incremental steps makes complex tasks less intimidating.
• Weaknesses: The bar for achieving mastery is too low and provides a false sense of security; the agents should query established content sources to guide analysis. Writing free-form learning objectives is confusing; templates and AI-suggested tasks should be included.