Edumimic

A smart teaching simulator that helps educators practice, improve, and engage effectively.

Comment

EduMimic - AI-Powered Teacher Training Simulator

🛠️ Hackathon Track

Tools for Teachers | AI in Education

Inspiration

Teaching can be challenging, especially for new educators who lack real-time feedback. I wanted to create a virtual environment where teachers can practice lessons, engage with AI-generated students, and get actionable insights—making teacher training more effective, engaging, and futuristic.

What it does

EduMimic is a virtual teaching simulator that allows teachers to:

  • Conduct lessons with real-time speech transcription using Whisper API.
  • Interact with AI-generated students who ask questions and react dynamically based on lesson context.
  • Monitor student engagement through facial expressions and voice clarity.
  • Review session summaries including transcripts, emotions, AI interactions, and engagement scores.

It simulates a real classroom so teachers can practice, experiment, and improve without needing a live audience.

How we built it

I developed EduMimic solo, under tight hackathon deadlines, using:

  • Frontend: React + TypeScript + Vite for a fast, responsive interface.
  • Speech Transcription: Whisper API captures teacher speech in 10-second chunks for accurate transcripts.
  • AI Students: Gemini 2.5 Flash Lite API generates intelligent, context-aware student questions and reactions.
  • Facial Expression Analysis: face-api.js tracks emotions like happy, neutral, sad, and surprised, calculating real-time engagement scores.
  • UI & Components: Tailwind CSS + shadcn/ui for polished panels and intuitive controls.
  • Audio Handling: MediaRecorder API captures audio efficiently while skipping silent/noisy chunks.

Everything is integrated to provide a full teaching session experience with actionable feedback.

Challenges we ran into

  • Solo development: Managing AI, audio/video processing, and engagement analytics alone.
  • Tight deadlines: Discovered the hackathon late, requiring rapid prototyping and prioritization.
  • Dynamic AI classroom: Making AI students respond realistically to teacher questions was technically challenging.
  • Real-time analytics: Balancing smooth video/audio capture with emotion detection and AI responses.

Accomplishments that we're proud of

  • Built a fully functional virtual teaching simulator that feels like a real classroom.
  • Integrated speech transcription, AI student interactions, and engagement scoring in real-time.
  • Achieved dynamic, context-aware AI responses without any pre-training.
  • Designed a polished, interactive UI that is intuitive and teacher-friendly.

What we learned

  • Real-time audio and video pipelines in the browser are feasible and scalable.
  • AI-generated interactions can effectively simulate realistic classroom scenarios.
  • Combining facial expression analysis and voice clarity provides quantitative engagement insights.
  • Solo hackathon work teaches rapid prototyping, technical versatility, and focus under pressure.

What's next for EduMimic

  • Add more AI student personalities and classroom dynamics for richer interactions.
  • Implement advanced engagement metrics including voice clarity and attention tracking.
  • Expand session analytics with detailed insights for teachers.
  • Explore gamified feedback and progress tracking to make teacher training even more immersive.
Share this project:

Updates