-
-
Landing Page
-
Teacher Dashboard
-
Edit Course Page
-
AI-Suggested Video Sections
-
Edit Training Video Details
-
Affiliate Textbook to Course 1
-
Affiliate Textbook to Course 2
-
Chunks in Textbook Associated with Course
-
Teacher's Profile
-
Course Page (From Student's POV)
-
Quiz (From Student's POV)
-
Quiz Completion (From Student's POV)
-
Generated Course Reading From Transcript & Textbook Chunks
About MotorMind
Inspiration
MotorMind was inspired by the RMS Diagnostics Hackathon Challenge, which asked us to build tools that transform how automotive diagnostics is taught and learned.
One problem stood out immediately: automotive educators and content creators already spend hours researching, filming and editing genuinely educational videos, only to repeat much of the same work creating online learning materials afterwards.
We asked a simple question:
Why should educators have to teach twice?
If someone has already taught something well in a video, AI should be able to turn it into a structured learning experience automatically.
What it does
MotorMind transforms automotive training videos into interactive learning experiences.
Upload a training video and MotorMind can automatically generate:
- Structured lesson sections from timestamps
- AI-generated study notes and reading
- Textbook-backed learning using semantic search
- Auto-generated quizzes
- Diagrams and visual explanations
- A conversational AI tutor students can speak to
- Progress tracking and achievement badges
Students can ask questions naturally, revisit key moments in videos, and receive explanations grounded in both the transcript and trusted textbook content.
How we built it
We built MotorMind as a full-stack web platform using:
- Google Gemini for educational content generation, quiz creation and tutoring
- ChromaDB vector search for textbook retrieval (RAG)
- ElevenLabs for conversational voice tutoring
- Django + SQLite for the platform backend
- Solana Devnet for verifiable achievement badges
To keep content reliable, we grounded AI responses using:
- video transcripts,
- textbook chunks retrieved through similarity search,
- citations with timestamps and page references.
This helps reduce hallucinations and makes the tutor feel more trustworthy in a vocational setting.
Challenges
The biggest challenge was balancing automation with accuracy.
Automotive diagnostics is practical and technical — incorrect information can be genuinely misleading. We therefore focused heavily on citations, transcript grounding and textbook retrieval.
Another challenge was ensuring teachers stayed in control. AI can generate materials quickly, but educators still need the ability to edit quizzes, readings and course content before publishing.
What we learned
We learned a huge amount about:
- retrieval-augmented generation (RAG),
- conversational AI,
- educational UX,
- grounding AI for technical subjects,
- and turning existing educational videos into scalable learning experiences.
Log in or sign up for Devpost to join the conversation.