Zero to DevOps+AI Ready In 90 DAYS

The complete beginner's DevOps roadmap for 2026, with free courses, crash courses, and an AI-first workflow.

Author

Vishakha Sadhwani
August 20, 2026

This is the only 90-day plan you need to go from zero to DevOps, with AI layered on top. The commitment is 6 hours a week. That's it.

The schedule is simple:

  • Mon / Wed / Fri: one 90-minute session each, one topic per session, nothing else open.

  • Tue / Thu / Sat: no new material. Redo the previous day's topic so it sticks.

  • Total: 3 new-material sessions a week, ~6 hours.

Your job here is to deploy and run applications, not write them. Keep that framing in mind through all three months. Below is every topic in the plan, why it matters, and exactly where to learn it for free.

Month 1: Core Foundations

You're not starting with system design. You're starting with the ground everything else stands on: the operating system, networking, version control, a little Python, and an AI assistant wired into your workflow.

Operating System (Linux or Windows)

Almost every server, cloud instance, and container runs on Linux, so you need to be comfortable at the terminal. Understand how a process works, what a service is, how memory gets used, and how permissions decide who can do what. If your target roles are Windows-heavy, learn the equivalents there, but Linux is the safer default.

Free courses

Networking

Before your code ever runs, a request travels through DNS, hits a port, and usually passes through a proxy like NGINX or a load balancer first. Understanding that request flow, and what sits in front of your application, is what separates someone who can deploy from someone who just writes code.

Free courses

YouTube videos

Git

Git is the version control system behind virtually all modern software. Before you touch any CI/CD pipeline, you need the commands, the workflow, and the lifecycle down cold: commits, branches, merges, and resolving conflicts.

Free courses

Blogs / docs

YouTube videos

Python

You are deploying the application, not writing it. So you don't need to master Python, you need to understand how an app is laid out: the project structure, what requirements.txt does, and how dependencies get installed. If you want to push into automation and scripting this month, go for it, but the baseline is reading and shipping, not building.

Free courses

AI Coding Assistant (the differentiator)

This is the most important habit of the month. Don't just use an AI assistant, wire it in properly: connect it to MCP (Model Context Protocol) servers and use skills so it can act on your tools and context, not just chat. Integrate AI into as much of your workflow as possible from day one.

Free courses / docs

YouTube videos

Tip: pair Cursor or Claude Code with an MCP server for something you actually use (GitHub, your cloud, a docs source) so the assistant becomes part of your real workflow, not a toy.

Month 2: Core Stack

Here the order matters. You move from a single cloud platform up through containers, orchestration, and infrastructure as code, and you finish the month reading how the giants built their systems.

Cloud Platform (AWS / GCP / Azure)

Pick one cloud and familiarize yourself with the core components, just the ones that help you deploy and serve application code. You don't need every service. You'll pick up the rest when you build an end-to-end project. Watch the crash course for the map, then the full course for depth.

Free courses

Docker

Docker popularized containerization, packaging an app with everything it needs so it runs the same anywhere. Learn it practically: multi-stage builds, container networking, and container scanning. Basically, how to containerize an app and drop it into a pipeline.

Free courses

YouTube videos

Kubernetes

Kubernetes is extremely industry-relevant right now and looks great on a resume, so build this skill however you can. Focus on scheduling, scaling, and networking, basically why Kubernetes exists. K8s knowledge is gold at the moment, and that's not an exaggeration.

Free courses

Blogs / docs

YouTube videos — a 3-part series to learn it end to end:

Terraform (+ your end-to-end project)

Terraform is the leading Infrastructure as Code tool. Watch one full course, then build an end-to-end project so you understand how a full pipeline gets deployed and how state management works. By the end you should be able to read Terraform files, write them, manage state, and see how it fits into a DevOps pipeline.

Free courses / docs

Build this end-to-end Terraform project (do it inside these 3 months):

Case Studies (system design in action)

This is the reading-heavy part of the month. Study how companies like Netflix, Uber, and YouTube build large systems. This is system design in action, and it gives you the "why" behind every tool you just learned.

Free resources

YouTube videos

Month 3: Standing Out

This is the month that separates you. A little AI infrastructure knowledge, one full-stack AI app deployed and troubleshot, a foot in the open-source world, and optionally a certification path to keep you on track.

AI Basics & AI Infrastructure

A short read every day. You are not building models, you're learning the difference between traditional infrastructure and AI infrastructure. Understand what a model is, how it runs, the key workloads (training vs inference), and how a model gets deployed.

Free resources

Full-Stack AI Application

Deploy a full-stack AI app and trace how a request flows from the open internet into your application. You'll touch API gateways, load balancers, and north-south vs east-west traffic patterns, and you'll troubleshoot a lot. Spend real time on troubleshooting, that's what makes you valuable.

YouTube videos

Supporting crash courses

Get Connected to Open Source

Pick one open-source project close to the job you want and get plugged in through its GitHub, Slack, Discord, or Telegram channels. Read what they're building. You won't understand much at first, keep reading anyway. This is how you learn the real, current state of the field.

Where to start

Certifications

If a certification learning path helps you stay disciplined, this is the time to sign up. It's a scaffold, not the goal.

Options

Crash Course Quick List

Every crash course video in one place, in case you just want the links:

Conclusion

Ninety days, six hours a week. At the end you won't have a certificate, you'll have something better: something that actually runs, and a clear idea of what you want to specialize in next.

Pick one topic per session. Do the reading, then get hands-on. Break things, fix them, repeat. That's the loop that builds a real DevOps engineer, and layering AI on top of it is what makes you stand out.

All links above are free courses, crash courses, official docs, or free-to-follow projects. Certification names, prices, and exam versions change, so confirm details on the official pages before booking.