AI-Powered Personalized Learning Assistant with Smart Content Chatbot

[sinegalatha]

Project Title

AI-Powered Personalized Learning Assistant with Smart Content Chatbot

Overview of the Idea

This project is an intelligent, end-to-end personalized learning platform designed to assist users in achieving their learning goals efficiently. Key features include:

• User Registration & Profiling: Captures personal (name, email, phone, age) and technical (skills, learning goals, preferred topics, and desired proficiency levels) information to build a learner profile.
• Smart Content Recommendation Engine: Automatically curates the best online learning resources (tutorial websites and YouTube videos) tailored to the user’s interests and expected proficiency.
• Dynamic Learning Assistant: Users can request content recommendations for any topic and skill level, receiving high-quality, publicly available learning materials.
• Admin Dashboard: Enables administrators to manage multiple user profiles, retrieve personalized learning content, and share resources via email or any platform.

Interactive Chatbot for Content Understanding:

• YouTube Video Chatbot: After selecting a video, users can interact with a chatbot to ask topic-specific questions and get accurate answers based on the video’s content.
• Document Chatbot: Similar chatbot interaction is available for selected PDF or Word documents, enabling a deep understanding of any textual study material.

Project Goal

To build an AI-powered personalized learning assistant that captures user profile details and recommends the most relevant online learning resources. It also provides intelligent chatbot support on top of YouTube videos and document files (PDF/Word) to make self-paced learning more effective and interactive.

How It Works

User Flow

User Registration:

• Users enter personal details and skills via Streamlit UI forms.
• Data is securely stored in a SQLite database.
• Smart Learning Recommendation:
• Based on user goals, the AI agent fetches tailored online learning resources (YouTube videos and public tutorials).

Admin Panel:

• Admin can register multiple users.
• Admin selects any user and retrieves personalized content suggestions, which can be shared via email or other platforms.

Video & Document Chatbot:

• Users can select a YouTube video or upload documents.
• An AI chatbot is enabled to answer questions based on the selected video or file content.

Core Functionality

User Profile Management: Collects personal and technical learning data.
AI-Powered Recommendations: Recommends best-suited YouTube videos and tutorials.
Admin Control: Register/view users and distribute learning resources.
Video Chatbot: Conversational AI trained on YouTube video transcripts.
Document Chatbot: Conversational AI trained on uploaded Word/PDF content.

Multimodal Elements* (if applicable):

Input Type	Usage
Text (User Profile & PDF/Word)	Used for recommendations and chatbot input.
Audio (YouTube Videos)	Extracted via moviepy and transcribed using Whisper model.
Video (YouTube)	Transcripts analyzed and indexed for chatbot interaction.
Chat Interface	Interactive chatbot embedded in Streamlit for real-time queries.

Tools Used

Purpose	Tool / Library
UI Framework	Streamlit
Database	SQLite
AI Agent	Agno Agent
LLM Model for Recommendations	AzureOpenAI - GPT-4.0
LLM Model for Document Chatbot	AzureOpenAI - GPT-3.5 Turbo
Embedding Model	AzureOpenAIEmbeddings
Vector Store (Document Chatbot)	FAISS
Video Processing (Video Chatbot)	moviepy - VideoFileClip
Speech to Text (Video Chatbot)	speech_recognition + Whisper
Vector Store (Video Chatbot)	VectorStoreIndex

UI Approach

• Built entirely using Streamlit for rapid development and clean layout.
• Responsive forms for user registration.
• Dropdowns for admin user selection.
• Embedded chatbot interfaces for video and document Q&A

Visuals

My Profile tab provides a comprehensive overview of the user's learning profile:
Left Pane: Displays the user's personal details, skills, learning goals, and the topics they are interested in.
Right Pane: Presents personalized learning recommendations generated based on the user’s selected goals and topics of interest.

The Admin Panel contains two key functionalities:
Register Profile
Profile-Based Recommendations

Register Profile: This section allows the admin to register new users by entering their personal details, skills, learning goals, and topics of interest. It helps populate the system with meaningful user profiles for personalized learning recommendations.

Profile-Based Recommendations: In this section, the admin can select a user from the list of registered users and view personalized learning recommendations. These are automatically generated based on the user's defined learning goals and topics, helping guide their educational journey by sharing it through email or in any platform.

In the Topic-Based Recommendation tab, users can:
- Enter any topic they wish to learn.
- Select their expected proficiency level (e.g., Beginner, Intermediate, Advanced).
Based on the input, the Recommendation Agent intelligently suggests relevant learning resources tailored to the selected topic and desired proficiency level, helping users get targeted and effective learning material.

Under the Chatbot with Resources tab, users will find two sections:
Video-Based Chatbot
- Users can browse the available list of videos and select any video to begin.
- Once the video starts playing, a chatbot specific to that video will be activated within a few seconds.
- Users can then ask questions related to the video content, and the AI agent will provide accurate and context-aware answers.

Document-Based Chatbot
- Users can browse and select any document (PDF) from the available list.
- A chatbot for the selected document will be activated shortly.
- Users can ask questions about the document content, and the AI agent will respond with relevant and precise answers.

Team Information

• Team Members:
  Sinegalatha B [sinegalatha2001@gmail.com] - VideoChatbot tool, DocumentChatbot tool
  Srikartikeya Kalluri[kartikalluri123@gmail.com] - Recommendation tool, Frontend UI design

Prize Category (leave blank, to be assigned by judges)

• Best use of Agno
• Best use of Firecrawl
• Best use of Mem0
• Best use of Graphlit
• Best use of Browser Use
• Best use of Potpie
• Best Overall Project

Demo Video Link

DemoVideoOfApplication.webm

Additional Notes

• Designed for scalability with modular AI services.
• Capable of supporting multiple users and customized learning paths.
• Ensures content relevance by allowing users to query AI on uploaded materials.
• Future enhancements may include:
  • Progress tracking.
  • Support for more file formats.
  • Integration with LMS platforms or calendar reminders.