Generative AI / Education
PrimaryTeachingSMA.
A multi-agent assistant for creating teaching exercises and encouraging computational thinking.
Problem
Help primary-school teachers prepare exercises that support computational thinking.
Context
Teaching assistant combining content generation, document retrieval and PDF export.
Objective
Orchestrate exercise creation with retrieved document context and PDF output.
Dataset / Data
Document context through ChromaDB. Corpus composition, rights and evaluation remain to be confirmed.
Architecture
Streamlit → CrewAI workflow → LangChain / RAG / ChromaDB → Claude → ReportLab.
Methodology
Separate agent, RAG, workflow and PDF modules: app.py, agents.py, rag.py, workflow.py, pdf_generator.py and test_api.py.
My contribution
Project reported in the candidate’s background. Individual scope and team responsibilities remain to be confirmed.
Challenges
Exercise consistency, retrieved context quality and cost control.
Limitations
No measured educational improvement is claimed. Exercises require teacher review.
What I learned
Practice in multi-agent workflows, RAG and document generation.
Project results
Reported average cost / run
Project-specific approximation. Model, pricing and methodology need confirmation.
Reported PDF generation
Approximate time under project conditions.