New: Roadmaps ordered paths through our cheat sheets and flashcards, so you always know what to study next.
Explore themSee what's new on GitHubFrom pay-as-you-go basics to a well-architected AWS deployment.
A 16-step learning path. Follow it in order, or jump to what you need.
This path is for anyone who wants to run systems on AWS, Azure, or GCP and know exactly what they're paying for, whether you're coming from on-prem IT, a junior dev role, or a fresh start in tech. Plan on about 8 to 10 weeks at 3 to 5 hours a week, starting from what cloud computing even means and ending at a deployment you could defend in a design review. It goes deep on the vendor platforms and services themselves, not on the CI/CD pipelines and team practices around them (see the DevOps Engineer path) or on going deep on one infrastructure-as-code tool (see Infrastructure as Code with Terraform). By the end you can stand up a secure VPC with locked-down IAM policies, pick the right compute, storage, and database service for a workload, and review an architecture against the AWS, Azure, and GCP Well-Architected pillars.
No prior experience needed. Start from zero.
Open with this and you'll leave knowing why IaaS, PaaS, and SaaS aren't just buzzwords, the vocabulary step 2 uses immediately to sort real instance types into their right category.
Builds on the service models from step 1 by putting real instance families and pricing levers in front of you, so step 3's console choices stop feeling arbitrary.
Takes the generic compute picture from step 2 and grounds it in one real provider's console, the platform most cloud job postings name first.
This is where VPCs, subnets, and CIDR math finally click for most people; expect to come back and reread it once you've actually built a network, since every workload step after this one assumes you can draw one.
You can explain what you're renting from a cloud provider in plain terms and sketch a VPC that keeps its pieces from talking to strangers. Next up: locking down who's actually allowed to touch any of it.
Finish this section to unlock.
+100 XP
Permissions trip up almost everyone the first time, and getting comfortable with least-privilege policies here is what keeps step 6's storage buckets from ending up public by accident.
Puts the access-control habits from step 5 to work, choosing between object, block, and file storage for whatever you're about to build.
Goes deeper on the data layer than step 6's storage classes, matching relational, document, and vector workloads to the right managed service.
Combines the compute instincts from step 3 with the execution roles from step 5 into functions that run without you managing a single server.
Shows the other end of the workload spectrum from step 8's functions, comparing how EKS, AKS, and GKE hand off cluster operations you'd otherwise own yourself.
You can lock a bucket down to least privilege and choose between a managed database, a serverless function, and a Kubernetes pod for wherever your workload actually needs to run. A few minutes of due flashcards keeps the VPC habits from chapter one fresh while you build on them here. Next up: making sure traffic reaches what you built, and what it costs to keep running.
Finish this section to unlock.
+100 XP
Ties the compute options from steps 3, 8, and 9 together by routing real traffic to whichever of them you picked.
Bound for a Microsoft-heavy team? This maps the same compute, storage, and IAM ideas from steps 3 through 5 onto Azure's resource hierarchy instead of AWS's.
Heading toward an AI-forward or GCP-first shop, this reuses the core services logic from step 3, now organized around Google's project hierarchy.
Every service you've provisioned since step 3 shows up on a bill, and this is how you keep that bill from becoming the reason your project gets shut down.
You can route traffic to a live service, recognize the same core ideas on Azure or GCP if a job ever calls for it, and catch a runaway bill before finance does. Next up: turning all of this into code you can hand off instead of clicks only you remember.
Finish this section to unlock.
+100 XP
Rebuilds everything from steps 3 through 10 as version-controlled config instead of console clicks, so your setup survives being handed to a teammate.
Want one IaC tool in real depth beyond step 14's concepts? State, modules, and HCL syntax get their own full workout in the dedicated Terraform path.
Pulls every pillar you've touched, security from step 5, cost from step 13, reliability from step 10, into the single framework AWS, Azure, and GCP all use to judge whether an architecture is production-ready.
You can provision a secure, multi-service cloud environment as version-controlled code and defend every choice in it against the Well-Architected pillars: security, cost, reliability, and performance. That's the Cloud Engineer badge, earned.
Finish this section to unlock.
+100 XP
Finish every required step, at least 70% of them genuinely done (not skipped), to earn this badge and 500 XP.