Skip to main content

Project Details

Physical AI & Humanoid Robotics Interactive Textbook

An interactive learning platform for Physical AI and Humanoid Robotics, built using AI-assisted development and spec-driven methodology.

Live Deployment​

Platform Features​

Comprehensive Interactive Textbook​

  • 10 Detailed Chapters covering Physical AI fundamentals, ROS 2, simulation environments, perception systems, SLAM, manipulation, deep learning, control systems, and deployment
  • Mobile-responsive design optimized for all devices
  • Full bilingual support (English/Urdu) with RTL layout
  • Professional Docusaurus-based documentation framework

Intelligent RAG Chatbot​

  • Context-aware Q&A powered by GPT-4 and vector search
  • Text selection support for targeted questions
  • Streaming responses with markdown formatting
  • Conversation history and persistence
  • Automatic semantic retrieval from textbook content

Smart User Profiles​

  • Detailed background assessment during registration
  • Automatic skill level detection (beginner/intermediate/advanced)
  • Profile management with editable preferences
  • Session persistence across devices

AI-Powered Content Personalization​

  • One-click content adaptation to user skill level
  • Dynamic content rewriting based on background
  • Personalized content saved per user, per page
  • Cross-session persistence

Personal Note-Taking System​

  • Page-specific notes for every chapter
  • Rich text editing capabilities
  • User-scoped note management
  • Persistent storage with easy deletion

Language Internationalization​

  • Complete English and Urdu translations
  • RTL (right-to-left) layout support for Urdu
  • Seamless language switching
  • All UI elements fully translated

Technology Stack​

Frontend Architecture​

  • Docusaurus 3.x - Modern documentation framework
  • React 18 - UI component library
  • TypeScript - Type-safe development
  • React Markdown - Formatted chatbot responses
  • Vercel - Edge deployment and hosting

Backend Services​

  • FastAPI - High-performance Python API framework
  • PostgreSQL (Neon) - Serverless database for user data and conversations
  • Qdrant Cloud - Vector database for semantic search and RAG
  • OpenAI GPT-4 - Language model for AI responses and personalization
  • Render - Cloud backend hosting

AI & Development Methodology​

  • Claude Code (Sonnet 4.5) - AI-assisted development partner
  • Spec-Kit Plus - Spec-driven development framework
  • RAG Architecture - Retrieval-Augmented Generation for accurate responses
  • Git + GitLab - Version control and collaboration

Libraries & Tools​

  • bcrypt - Secure password hashing
  • PyJWT - JWT-based authentication
  • SQLAlchemy - Database ORM
  • python-dotenv - Environment configuration
  • CORS Middleware - Cross-origin security
  • React Router - Client-side navigation

Development Approach​

This project was built using Spec-Driven Development methodology:

Planning & Documentation​

  • Feature Specifications - Detailed requirements and user stories documented in /specs
  • Implementation Plans - Architectural decisions and design patterns in plan documents
  • Task Breakdown - Granular, testable tasks with acceptance criteria
  • Architecture Decision Records (ADRs) - Key technical decisions documented with rationale

AI-Assisted Development​

  • Explore Agent - Automated codebase analysis and discovery
  • Plan Agent - Implementation planning and architecture design
  • Task Agent - Multi-step execution and automation
  • Prompt History Records (PHRs) - Complete development history tracking

Quality Artifacts​

All development documentation included:

  • /specs - Feature specifications and plans
  • /history - PHRs and development timeline
  • /.specify - Spec-Kit Plus templates and scripts
  • Comprehensive guides and methodology documentation

Educational Content​

Textbook Curriculum​

  1. Introduction to Physical AI - Foundations and core concepts
  2. ROS 2 Fundamentals - Robot Operating System essentials
  3. Simulation Environments - Gazebo, Isaac Sim, and virtual testing
  4. Perception Systems - Sensors, vision, and environmental awareness
  5. SLAM & Navigation - Simultaneous localization and mapping
  6. Manipulation & Grasping - Robotic arm control and object handling
  7. Deep Learning for Perception - Neural networks for robotics
  8. Control Systems - Motion planning and control theory
  9. Advanced Topics - Cutting-edge research and techniques
  10. Physical Deployment - Real-world implementation strategies

Plus comprehensive methodology guides and development workflows.

Core Capabilities​

✅ Interactive Learning Platform - Complete textbook with navigation and search ✅ RAG-Powered Q&A - Intelligent chatbot answering questions from textbook content ✅ Text Selection Queries - Ask questions about specific highlighted content ✅ User Authentication - Secure signup/signin with JWT tokens ✅ Adaptive Personalization - AI-powered content adjustment to skill level ✅ Multilingual Support - Full English and Urdu translation ✅ Cloud Deployment - Production-ready on Vercel and Render ✅ Spec-Driven Process - Complete documentation and planning artifacts

Design Philosophy​

  • Minimalistic - Clean, focused interface without clutter
  • Professional - Modern, polished visual design
  • Accessible - Clear navigation and readable typography
  • Responsive - Seamless experience across all screen sizes
  • Intuitive - Features are easy to discover and use

Architecture Highlights​

Backend API Structure​

/api/auth         - User authentication (signup, signin, profile)
/api/chat         - RAG chatbot endpoints (query, conversations)
/api/personalize  - Content personalization (adjust, save, reset)
/api/admin        - Database management (migrations)

Database Schema​

  • Users - Authentication, profiles, skill levels
  • Conversations - Chat history with titles and timestamps
  • Messages - Individual chat messages with roles
  • PageEdit - Personalized content per user/page

Vector Search Pipeline​

  1. Content chunking from textbook markdown
  2. OpenAI embeddings generation (1536 dimensions)
  3. Qdrant vector storage with metadata
  4. Semantic similarity search
  5. Context-aware response generation

Contact & Resources​

Developer: Ahtisham Yousaf Repository: GitLab - physical-ai-textbook Live Platform: physical-ai-textbook.vercel.app

Built with Claude Code and Spec-Kit Plus methodology.

An AI-powered educational platform demonstrating modern development practices and intelligent learning systems.