Goal: <your target cert, e.g. GCP Associate Cloud Engineer>, ready to apply in 3 months (from , targeting ). Self-learning pace, commitment.
<your target role, e.g. Platform / DevOps Engineer (mid-level)> in <your city / region>. Study 10-15 real job ads from your market and list the stack they actually ask for. Typical stack: Linux, Docker, Kubernetes, Terraform/Ansible, Jenkins / GitLab / GitHub Actions, AWS or GCP, Prometheus+Grafana or Datadog.
- Weeks 1-3: Linux on WSL2 / Ubuntu VM (Linux Journey, freeCodeCamp). Filesystem, permissions, users/groups, processes, systemd, package managers, bash scripting.
- From week 3 onward: write daily automation scripts in Python as well as bash — JDs ask for "Python, Go, or Shell", so both get reps every week.
- Week 4: Networking fundamentals (OSI, TCP/IP, DNS, HTTP, ports, firewalls, subnets, load balancing). Free: Professor Messer Network+ videos.
- Milestone: pass LPI Linux Essentials.
- Week 5: Docker deep-dive - multi-stage builds, networks, Compose.
- Weeks 6-7: Kubernetes - minikube/Kind locally, then a managed cluster (GKE/EKS/AKS). Pods, deployments, services, ingress, configmaps, secrets, storage.
- Week 8: Terraform (IaC) - re-create one of your existing deployments as code.
- Milestone: running K8s project on an existing app.
- Week 9: Prometheus + Grafana dashboard; deepen GitHub Actions (environments, CD). One afternoon: GitOps — install ArgoCD on the K8s cluster and deploy by git push.
- Weeks 10-11: cloud cert prep + exam (default: GCP ACE — swap for your cloud).
- Week 12: study 10-15 DevOps/Platform JDs from your target market, extract exact keywords, close remaining gaps, log new achievements to your resume, tailor applications.
Run by the /coach skill: /coach in the morning for the day's tasks,
/coach done in the evening for the proof check. State lives in progress.md,
daily notes in notes/.
- 2h hands-on first, best energy. Terminal, not videos. Break things, fix them.
- 1h new material - video/reading for the current week's topic.
- 30m write it down - one note per day: what I did, what broke, how I fixed it. Feeds the portfolio and future resume bullets.
- 30m small reps - 10m review yesterday's note, 10m cert practice questions, 10m read one DevOps JD from your target market and list unknown words.
- 15m ship to git - even a tiny script. Daily commits show the gap was worked.
Rules:
- Terminal only, no GUI. Type every command, never copy-paste while learning.
- One "why" per day: explain one thing out loud in simple words. If I can't, I don't understand it yet.
- Friday = rebuild day. Delete the week's setup, rebuild from scratch, faster.
- Weekend lighter: one 2h session, then rest. Burnout in week 5 kills the plan.
- Daily question: "What can I do today that I couldn't do yesterday?"
- Self-hosted stack at home: reverse proxy + Docker Compose + Prometheus/Grafana + CI/CD.
- AI serving on K8s: <one of your own apps, ideally with an LLM backend> on a real K8s cluster — resource limits, probes, metrics + cost-per-request in Grafana, infra in Terraform. This is the "AI-platform" proof piece.
- IaC reproducibility: full stack destroyed and rebuilt from one command.
- Week 1-3: Linux daily practice done
- Week 4: networking fundamentals done
- Week 4: LPI Linux Essentials passed
- Week 5: Docker deep-dive done
- Week 6-7: K8s project running
- Week 8: Terraform IaC done
- Week 9: Prometheus/Grafana live
- Week 10-11: cloud cert passed
- Week 12: JDs studied, gaps closed, applications tailored
- Never fabricate achievements; only log real results to your resume.
- Later: add CKA once employed, then consider remote/international market.