https://github.com/anis-marrouchi/genai-curriculum-demo
https://github.com/anis-marrouchi/genai-curriculum-demo
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/anis-marrouchi/genai-curriculum-demo
- Owner: anis-marrouchi
- Created: 2025-07-18T17:30:37.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-07-18T19:30:04.000Z (4 months ago)
- Last Synced: 2025-08-08T22:48:13.124Z (3 months ago)
- Language: Python
- Size: 20.5 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AutoGen Curriculum Demo
Multi-agent curriculum generation using Microsoft AutoGen - showcasing conversational AI
agents collaborating to create structured educational content. Intelligent, coordinated,
and effective AI-powered curriculum development.
## Quick-start
```bash
git clone https://github.com/anis-marrouchi/genai-curriculum-demo.git
cd genai-curriculum-demo
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env # add your OpenAI key
streamlit run src/app_streamlit.py # or python src/run_demo.py
```
## Approach
This demo uses Microsoft AutoGen's multi-agent framework with specialized agents following a structured workflow:
- **Sequential Processing**: Idea → Objectives → Lessons → Assessments
- **Quality Review Loop**: Automated peer-review with feedback routing
- **State Management**: Tracking workflow progress and iterations
## Agents
- **ObjectivesExpert**: Creates SMART learning objectives using Bloom's taxonomy
- **LessonDesigner**: Builds detailed lesson blueprints with HOOK-EXPLAIN-PRACTICE-REFLECT model
- **AssessmentCreator**: Generates aligned assessments with MCQs and rubrics
- **QualityReviewer**: Reviews curriculum alignment and provides feedback for iterations
## Structure
```
genai-curriculum-demo/
├── README.md
├── .env.example
├── requirements.txt
├── src/
│ ├── __init__.py
│ ├── llm.py # AutoGen configuration factory
│ ├── state_schema.py # Workflow state management
│ ├── workflow.py # Multi-agent workflow coordinator
│ ├── curriculum.py # Main curriculum generation interface
│ ├── agents/ # Individual agent implementations
│ │ ├── __init__.py
│ │ ├── idea_to_objectives.py
│ │ ├── objectives_to_lessons.py
│ │ ├── lessons_to_assessments.py
│ │ ├── critique.py
│ │ └── package.py
│ ├── run_demo.py
│ └── app_streamlit.py
```
## Benefits
- **Modular**: Each agent in separate file, following LangGraph pattern
- **Intelligent**: Specialized agents with expert personas and conversation handling
- **Coordinated**: Structured workflow with conditional routing and iteration loops
- **Robust**: Comprehensive error handling and fallback mechanisms
- **Trackable**: Full conversation history and workflow state management
- **Scalable**: Easy to add new agents or modify workflow steps