MLOps Platform Engineering Lab

Hands-on lab documenting my transition from Senior DevOps / Integration Engineer (12 years, telecom infrastructure — Jio, Globe, Etisalat, Batelco, Telenor, Ericsson) to AI Platform / MLOps Engineering. Every manifest here was built and validated on a local cluster, not copied from a tutorial.

Structure

• k1-core-objects/ — Kubernetes core objects: Deployment, Service, ConfigMap, label selector mechanics, reconciliation model
• k2-hpa-custom-metrics/ — Custom metrics autoscaling pipeline for an ML inference workload (Prometheus + Prometheus Adapter + HPA)
• k3-storage-rbac-networking/ — Persistent storage, RBAC, NetworkPolicy isolation (in progress)
• app/fastapi-app/ — FastAPI inference service instrumented with prometheus-client

Why this exists

Targeting AI Platform Engineer / MLOps Engineer roles (IC track) at GCCs in Delhi NCR. This repo is the proof-of-work layer behind that transition — 12 years of telecom-scale infrastructure experience applied to AI infra.

Environment

Windows + WSL2, Rancher Desktop (k3s), containerd runtime.