Docker

AI apps have complex dependencies. Docker solves 'it works on my machine' with reproducible environments.

A shipping container — loads and unloads the same way everywhere, no matter what's inside.

Visual Workflows

Start here — study each diagram, then use + / − to zoom if needed.

Overview

100%

Loading diagram...

Scroll inside the frame · use + / − to zoom

100%

Loading diagram...

Scroll inside the frame · use + / − to zoom

Key Takeaways

1.Docker packages your app and dependencies into a portable container — runs identically on your laptop and in production.
2.Images are blueprints, containers are running instances.
3.Dockerfile defines the build.
4.Docker Compose runs multiple services together.

Real Example

Scenario

Your FastAPI app works locally but crashes on the production server due to missing dependencies.

What you would do

Dockerize it — Dockerfile installs exact deps, docker-compose runs API + ChromaDB together. Same result everywhere.

Practice Task

Run docker run hello-world. Then run docker ps -a to see the container. If Docker is not installed, read the Docker diagram and write down the 4 key commands you will use later.

Commands

Commands to Remember

docker build -t name . # build an image from Dockerfile
docker run -p 8000:8000 name # run container and map port 8000
docker run --env-file .env name # pass secrets via env file at runtime
docker-compose up -d # start all services in the background
docker logs container_id # view logs from a container
docker ps # list currently running containers

Cheat Sheet

Quick recap

quick ref

•Image = blueprint · Container = running instance
•Dockerfile defines build steps · docker build creates image
•docker run starts a container · docker-compose runs multiple services
•Pass secrets via --env-file at runtime — never in Dockerfile
•Use .dockerignore to keep images small
•docker logs and docker exec for debugging

Common Mistakes

✕API keys in Dockerfile
✕No .dockerignore — huge images
✕Running as root in production