Docker Interview Questions for Freshers

Docker is a containerization platform that packages your application and all its dependencies into a container — a portable, isolated environment. For AI: (1) Reproducibility — same Python version, library versions everywhere, (2) Dependency isolation — no conflicts between projects, (3) Scalability — easily spin up multiple instances, (4) CI/CD integration — build once, deploy anywhere. Without Docker, "it works on my machine" kills AI deployments.

Docker Compose orchestrates multiple containers for local development. For an AI app stack: services: app (FastAPI), db (PostgreSQL), cache (Redis), vectorstore (Chroma). Define in docker-compose.yml with ports, volumes for hot reload, and environment variables. Use depends_on with health checks. In AI development, also add: a mock LLM service for testing without API costs, and Prometheus + Grafana for local monitoring.

Volumes are managed by Docker at /var/lib/docker/volumes — portable, backed up, shareable between containers. Bind mounts link a host directory into the container — used for hot reload in development (COPY . . becomes volume: ./:/app). For AI: use volumes for model weights and database files in production, bind mounts for development to avoid rebuilding the image on every code change.