Definition
Kubernetes (K8s) schedules Docker containers across nodes, restarts failed ones, scales replicas under load, and exposes services with stable networking.
It is the default for running microservices and ML inference at scale in production clouds.
In simple terms
Docker is one shipping container; Kubernetes is the port that manages thousands of containers — loading, routing, and replacing damaged ones automatically.
Where you see it
- SaaS backends run as K8s deployments.
- GPU inference servers scale with horizontal pod autoscaling.
- Managed K8s: GKE, EKS, AKS.
How it works
1.Define manifests
YAML describes pods, services, and deployments.
2.Schedule pods
Control plane assigns containers to nodes.
3.Expose services
Load balancers and ingress route traffic.
4.Scale and heal
Replica sets and health checks maintain desired state.
Why it matters
- K8s skills bridge DevOps and ML engineering for production AI services.
Often confused
Kubernetes is required for every app.
Small projects often use PaaS or single-container deploys; K8s pays off at scale and team size.