Speck

A robust and scalable platform designed to deliver personalized learning experiences through modern web technologies and AI integration.

Design and Development

We approached the development of Speck with a focus on creating a reliable and scalable system. Our carefully chosen tech stack includes:

Backend

• Node.js and Express
  Build a powerful and flexible server-side application.
  • Passportjs: For OAuth authentication
  • WebSocket Connection: Provides live updates for time-consuming services.
  • Redis: Efficient for caching and handling polling mechanisms.

Database

• PostgreSQL
  A robust database chosen for its capability to manage complex data relationships.
• Prisma
  An ORM that enables efficient and streamlined database operations.

Frontend

• Next.js
  Develops a responsive and performant user interface using the App Router, enhanced by the Shadcn component library.

Cloud Infrastructure

• Azure Services
  Ensures the scalability and reliability of the platform.

AI Integration

• Llama 3.3 and Langchain
  Power the AI-driven content creation and personalization features.

Project Structure

For local development, Docker was used to containerize the services. A Docker Compose file orchestrates multiple containers with proper networking, port mapping, and volume persistence.

Docker Compose Example

version: '3'
services:
  backend:
    build: ./backend
    ports:
      - "3000:3000"
    environment:
      - NODE_ENV=development
  frontend:
    build: ./frontend
    ports:
      - "3001:3000"
  redis:
    image: redis:latest
  postgres:
    image: postgres:latest
    environment:
      - POSTGRES_PASSWORD=yourpassword

Directory structure:

└── harshptl14-speck/
    ├── compose.yaml
    ├── speck-backend/
    └── speck-frontend/

Architecture Overview

The system is divided into key layers:

• Frontend: Interfaces with users.
• Backend: Hosts APIs and business logic, interacting with databases.
• Datastores: Includes both PostgreSQL and Redis.

API Design

The project includes several crucial APIs, particularly for course generation. Below is a breakdown of the key APIs:

1. Course Generation API

Add API

• Purpose: Processes the user's prompt and triggers the Content Generation service.
• Technologies: Utilizes Langchain and groq (Llama 3.3) for content generation.
• Real-Time Updates: Uses WebSocket to update the user on the progress of generating a complete topic (which may consist of multiple sub-topics).
• Reasoning: Suitable for longer processes where polling could lead to unnecessary API calls.

On-Demand Content Generation API

• Purpose: Generates content for a single sub-topic.
• Mechanism: Employs polling to monitor progress.
• Tracking: Uses a unique JobId stored in Redis to retrieve the latest progress updates.

Get API

• Purpose: Retrieves the generated data.
• Data Source: Fetches data directly from PostgreSQL or from Redis if available in cache.

Features and Functionality

Speck's core functionality centers around delivering personalized, bite-sized learning content. Key features include:

AI-Driven Content Creation

Automates the generation of tailored learning paths and micro-content using Llama 3.3 and Langchain.

Progress Tracking

Enables users to track their learning progress in a course-like view and mark topics as complete.

Relevant Visual Learning Content

For each topic and subtopic, the platform provides:

• Long and short-length videos
• Synchronized transcripts

Roadmap View

Offers a tree-like visualization that displays the complete roadmap, including topics and subtopics.

Mobile Optimization

Ensures seamless access to content on all devices, particularly smartphones, for on-the-go learning.

Conclusion

Speck represents a significant step forward in continuous learning. By integrating AI and modern web technologies, the platform delivers knowledge in a way that seamlessly fits into users' lives.

Note: The project is continuously evolving, with regular improvements and feature additions planned. Try it today at speck.ing