Docker for Local AI App Development: Build Lightweight, Containerized AI Applications

In this course for AI developers, learn how to make Docker containers part of the entire application lifecycle—from defining requirements and building a reproducible development environment to testing services and preparing images for production. Instructor Rami Krispin uses a retrieval-augmented generation (RAG) system as a running architectural case study, showing you how to identify application services, define their requirements, and choose container boundaries that fit both development and production needs.

Build a foundation in Dockerfiles, images, registries, and the build-and-run workflow; then use Docker Compose and VS Code Dev Containers to create a containerized workspace for prototyping and validating the application. See how to transform that prototype into dedicated ingestion, query, and vector database services, test the stack in an environment that closely resembles production, and prepare its images for release through multi-stage builds, security hardening, multi-platform builds, versioned publishing, runtime safeguards, and CI validation. By the end of the course, you’ll have a practical, reusable framework for developing, testing, and preparing multi-service AI applications for production at the image and container levels.

Learning objectives
Build and run AI applications entirely inside Docker.
Run lightweight local LLMs inside containers.
Integrate AI APIs from OpenAI and Anthropic inside a containerized app.
Build a containerized RAG prototype using an application container plus a vector store.
Debug and iterate on AI features inside containers.
Package and distribute a containerized AI microservice.