elkay:: readme

README.md

@@ -4,6 +4,16 @@
 $mart Kids is an AI-powered web-based learning platform designed to teach financial literacy through lessons, quizzes, and interactive games.
 Built with Streamlit, FastAPI, LangGraph, TiDB Cloud, and Hugging Face models, it leverages multi-agent AI workflows and vector search to provide personalized tutoring experiences for students, while giving teachers tools to manage classes and track progress.
 
+❓ Problem & Solution
+In Jamaica and many developing countries, students often lack engaging, accessible tools for learning financial literacy and numeracy. Traditional classrooms don’t adapt well to different learning paces, teachers struggle to track student progress at scale, and parents lack visibility.
+
+**$mart Kids solves these challenges by:**
+- Gamifying financial learning through XP and leaderboards.
+- Offering a personalized AI tutor that adapts to each student.
+- Providing teachers with dashboards, lesson creation tools, and real-time progress tracking.
+- Allowing parents to receive detailed reports on student progress.
+
+
 ✨ Features
 Student Dashboard
 Explore lessons, attempt quizzes (currently preloaded due to OSS constraints), receive AI tutor feedback, and play engaging games like Budget Builder and Debt Dilemma.
@@ -53,6 +63,27 @@ short_description: This is the Backend for Project - $mart Kids – AI Financial
 Built with FastAPI + LangGraph
 Handles agent flows, embeddings, and TiDB Cloud integration
 
+🛠️ Tech Stack
+Frontend: Streamlit (FinFront)
+Backend: FastAPI + LangGraph (FinBack)
+Database: TiDB Cloud (with Vector Search)
+Models: Hugging Face (TinyLlama, E5 embeddings)
+Deployment: Hugging Face Spaces
+
+
+🔄 Data Flow & Integration
+1. Student/Teacher Action → User interacts with FinFront (Streamlit).
+2. API Requests → FinFront communicates with FinBack (FastAPI).
+3. Database Operations → FinBack stores and retrieves lessons, quizzes, and user progress from TiDB Cloud.
+4. Vector Search → Lesson embeddings are stored in TiDB Cloud for semantic search and personalization.
+5. AI Models → Hugging Face models (TinyLlama for chat, E5 for embeddings) are used for tutoring, explanations, and adaptive learning.
+6. Feedback Loop → Results (XP, scores, progress) are sent back to dashboards in FinFront.
+
+
+📹 Demo Video
+👉 https://app.supademo.com/demo/cmflwngwdfpoh39ozaycikuib?utm_source=link
+
+
 🚀 Run Instructions
 Access Online
 1. Open FinFront on Hugging Face - https://finedproj-finfront.hf.space/
@@ -74,13 +105,6 @@ Teachers → lesson management, quiz tools, dashboards.
 (Note: quizzes are currently preloaded due to limitations in open-source LLM quiz generation.)
 
 
-🛠️ Tech Stack
-Frontend: Streamlit (FinFront)
-Backend: FastAPI + LangGraph (FinBack)
-Database: TiDB Cloud (with Vector Search)
-Models: Hugging Face (TinyLlama, E5 embeddings)
-Deployment: Hugging Face Spaces
-
 
 📜 License
 MIT License