Inspiration
Traditional learning systems treat all students the same, but every student learns differently. We wanted to build a system that understands each student individually.
What it does
Our AI Learning Digital Twin creates a personalized learning profile for each student. It tracks performance, identifies weak topics, predicts retention gaps, and generates quizzes, summaries, and improvement plans.
How we built it
We used Databricks for data processing (Delta Lake + SQL Warehouse), PySpark for analytics, and LLaMA 3.3 for AI features. A Streamlit frontend connects everything into an interactive app.
Challenges we ran into
Integrating real-time analytics with LLM outputs and managing data pipelines was challenging. Ensuring smooth frontend-backend communication also required debugging.
What we learned
We learned how to build end-to-end AI systems combining data engineering, analytics, and LLMs into a single product.
What's next
We plan to improve predictions, add adaptive learning paths, and scale this system for real classrooms.
Log in or sign up for Devpost to join the conversation.