Docker Workshop for Data Science & Machine Learning

A comprehensive hands-on workshop covering Docker fundamentals through deploying AI/ML applications to the cloud.

🎯 Workshop Overview

This workshop takes you from Docker basics to deploying a tiny LLM (Large Language Model) application in the cloud. Through practical exercises, you'll learn containerization concepts essential for modern data science and machine learning workflows.

📚 Workshop Modules

Module 1: Docker Basics

Duration: ~1 hour
Difficulty: Beginner

Learn the fundamentals of Docker containers and images.

What You'll Learn:

• Docker architecture and concepts
• Running containers from existing images
• Container lifecycle management (start, stop, restart)
• Port mapping and networking basics
• Working with container logs
• Interactive containers and shell access
• Resource monitoring and limits
• Environment variables
• Container cleanup and management

Key Exercises:

• Run your first Docker container (Hello World)
• Run multiple web servers on different ports
• Explore container logs in real-time
• Work with interactive Ubuntu containers
• Monitor resource usage with docker stats
• Clean up containers and images

---

Module 2: Building Docker Images

Duration: ~1.5 hours
Difficulty: Intermediate

Master the art of creating custom Docker images and learn Dockerfile best practices.

What You'll Learn:

• Dockerfile syntax and structure
• Building custom images from scratch
• Layer caching optimization
• Multi-stage builds (if applicable)
• Best practices for AI/ML workloads
• Security considerations (non-root users)
• Health checks for production
• Handling large dependencies (PyTorch, Transformers)

Key Exercises:

• Create your first Dockerfile
• Build a Story Generator TinyLLM application
• Optimize Dockerfile for layer caching
• Work with large ML dependencies
• Implement security best practices

Featured Project:

• Story Generator TinyLLM App
  • Uses DistilGPT2 (80MB lightweight LLM)
  • Streamlit chat interface
  • Optimized for 512MB RAM
  • Production-ready Dockerfile with health checks

---

Module 3: Docker Compose

Duration: ~1 hour
Difficulty: Intermediate

Learn to orchestrate multi-container applications with Docker Compose.

What You'll Learn:

• Why Docker Compose is needed
• YAML configuration syntax
• Service orchestration
• Docker networking between containers
• Volume management for data persistence
• Environment variable management
• Service dependencies and health checks
• Database initialization scripts
• Troubleshooting multi-container apps

Key Exercises:

• Set up a PostgreSQL + pgAdmin stack
• Configure Docker networks for service communication
• Persist data with Docker volumes
• Connect services using service names
• Test data persistence across container restarts
• Debug service communication issues

Featured Project:

• E-Commerce Database Demo
  • PostgreSQL database with sample data
  • pgAdmin web interface for database management
  • Automatic initialization with sample data
  • Data persistence with volumes

---

Module 4: Deploy TinyLLM to Cloud

Duration: ~1 hour
Difficulty: Intermediate to Advanced

Deploy your containerized LLM application to the cloud.

What You'll Learn:

• Cloud deployment pipeline
• Docker registries (Docker Hub)
• Image tagging strategies
• Pushing images to registries
• Cloud platform deployment (OnRender)
• Production considerations
• Getting a public HTTPS URL
• Sharing your application globally

Key Exercises:

• Create a Docker Hub account
• Tag and push images to Docker Hub
• Set up OnRender cloud platform
• Deploy from Docker Hub to cloud
• Access your live application via public URL

Deployment Pipeline:

Local Development → Build Image → Push to Docker Hub → Deploy to OnRender → Public URL

---

🚀 Getting Started

Prerequisites

• Docker Desktop installed (Download)
• Basic command line knowledge
• Text editor (VS Code recommended)
• Docker Hub account (for Module 4)
• OnRender account (for Module 4)

Quick Start

1. Clone or download this repository

   cd Docker-Workshop-DSML

2. Verify Docker installation

   docker --version
   docker run hello-world

3. Start with Module 1

   cd module-1-docker-basics
   # Follow exercises.md

---

📖 Module Structure

Each module contains:

• exercises.md - Step-by-step hands-on exercises
• Sample code and configuration files (where applicable)
• Expected results and validation steps
• Common errors and troubleshooting tips

Recommended Learning Path

1. Module 1: Docker Basics - Foundation concepts
2. Module 2: Building Images - Create custom containers
3. Module 3: Docker Compose - Multi-container apps
4. Module 4: Cloud Deployment - Go live!

---

🛠️ Tools & Technologies Used

• Docker - Container platform
• Docker Compose - Multi-container orchestration
• Python - Application development
• Streamlit - Web UI framework
• Transformers - Hugging Face ML library
• DistilGPT2 - Lightweight language model
• PostgreSQL - Relational database
• pgAdmin - Database management UI
• Docker Hub - Container registry
• OnRender - Cloud deployment platform

---

🎓 Learning Outcomes

By completing this workshop, you will be able to:

✅ Understand Docker architecture and container concepts
✅ Run and manage Docker containers
✅ Create custom Docker images with Dockerfiles
✅ Optimize images for AI/ML workloads
✅ Orchestrate multi-container applications
✅ Manage data persistence with volumes
✅ Configure container networking
✅ Deploy containerized applications to the cloud
✅ Share your work via public URLs
✅ Apply Docker best practices for production

---

🔧 Common Commands Reference

Basic Docker Commands

# Container management
docker run [IMAGE]              # Run a container
docker ps                       # List running containers
docker ps -a                    # List all containers
docker stop [CONTAINER]         # Stop a container
docker start [CONTAINER]        # Start a container
docker rm [CONTAINER]           # Remove a container
docker logs [CONTAINER]         # View container logs

# Image management
docker images                   # List images
docker build -t [TAG] .         # Build image
docker pull [IMAGE]             # Pull image from registry
docker push [IMAGE]             # Push image to registry
docker rmi [IMAGE]              # Remove image

# System management
docker system df                # Show disk usage
docker system prune             # Clean up unused resources

Docker Compose Commands

docker-compose up -d            # Start services in background
docker-compose down             # Stop services
docker-compose ps               # List services
docker-compose logs [SERVICE]   # View service logs
docker-compose restart [SERVICE] # Restart a service
docker-compose down -v          # Remove services and volumes

---

⚠️ Common Issues & Solutions

Port Already in Use

# Find what's using the port
docker ps

# Use a different port
docker run -p 8080:80 nginx  # Change 8080 to another port

Container Name Conflict

# Remove existing container
docker rm -f [CONTAINER_NAME]

# Or use a different name
docker run --name myapp-v2 nginx

Cannot Connect to Docker Daemon

• Ensure Docker Desktop is running
• On Linux: sudo systemctl start docker

Out of Disk Space

# Clean up unused containers and images
docker system prune -a --volumes

---

🎯 Projects Built in This Workshop

1. Story Generator TinyLLM App (Module 2)

• AI-powered story generation
• Uses DistilGPT2 model
• Streamlit web interface
• Optimized for low memory (512MB)

2. E-Commerce Database Stack (Module 3)

• PostgreSQL database
• pgAdmin web interface
• Sample e-commerce data
• Persistent volume storage

3. Cloud-Deployed LLM App (Module 4)

• Publicly accessible AI app
• HTTPS enabled
• Global deployment
• Production-ready

---

📝 Workshop Tips

For Beginners

• Take your time with Module 1
• Run each command and observe the output
• Read error messages carefully
• Use docker ps -a frequently to check container status

For Intermediate Users

• Focus on Dockerfile optimization in Module 2
• Understand networking in Module 3
• Experiment with different configurations

For All Levels

• Clean up containers regularly
• Check logs when things don't work
• Use docker inspect to debug
• Practice the commands multiple times

---

🌟 Next Steps After Workshop

1. Build Your Own ML App

   • Containerize an existing project
   • Try different ML frameworks (TensorFlow, scikit-learn)
   • Optimize for your specific use case

2. Explore Advanced Topics

   • Kubernetes for orchestration
   • Docker Swarm for clustering
   • CI/CD pipelines with Docker
   • Security scanning and best practices

3. Production Considerations

   • Container monitoring and logging
   • Auto-scaling strategies
   • Load balancing
   • Database backups and disaster recovery

4. Join the Community

   • Docker Community Forums
   • Stack Overflow
   • GitHub repositories
   • Local Docker meetups

---

📚 Additional Resources

• Docker Official Documentation
• Docker Hub
• Dockerfile Best Practices
• Docker Compose Documentation
• Hugging Face Transformers
• Streamlit Documentation

---

🤝 Contributing

If you find issues or have suggestions for improvement:

1. Open an issue
2. Submit a pull request
3. Share your feedback

---

📄 License

This workshop is provided for educational purposes. Feel free to use and modify for your learning and teaching needs.

---

🙏 Acknowledgments

Built with ❤️ for the Data Science and Machine Learning community. s Special thanks to:

• Docker community for excellent documentation
• Hugging Face for making ML accessible
• Streamlit for easy web app development

---

Ready to containerize your ML workflows? Start with Module 1! 🚀