
Four certification tabs open. AWS. Azure. Kubernetes. Some kind of AI credential. And one question: do you really need four separate plans?
No. You need one Cloud Certification Path โ a single roadmap where AWS Certification, Azure Certification, Kubernetes Certification, and AI Certification are checkpoints, not competing destinations.
Here's that roadmap: a 12โ24 month plan, a look at how much these four areas actually overlap, and sample paths by role โ so you stop guessing and start studying.
Most companies aren't single-cloud anymore. AWS here, Azure there โ often just because of mergers or history. Know only one platform, and you're only useful for half the conversation.
Kubernetes sits underneath most of it. EKS, AKS, self-managed โ pods and deployments work the same everywhere. Learn it once, work across clouds.
AI rides the same rails. Models get trained, served, and deployed as containerized workloads โ not isolated data-science side projects.
That's the case for one combined path instead of four:

Four pillars, one roof. Build them in sequence, not isolation.
โ ๏ธ Heads up: the old Machine Learning โ Specialty exam retired March 31, 2026. The current AI route is AI Practitioner โ ML Engineer Associate โ Generative AI Developer Professional.
โ ๏ธ Two exams are mid-transition: AI-102 retired June 30, 2026 (replaced by AI-103, built around agents and generative AI), and AZ-500 retires August 31, 2026 (replaced by SC-500). Study the current exams, not the outgoing ones.
Azure mirrors AWS closely enough that adding it second is much faster than starting cold.
CNCF runs five exams โ two multiple-choice, three hands-on:
This pillar is the bridge: the same skills apply whether you're running EKS or AKS. Once you're past the exam and running real workloads, MicroDegree's guide on Kubernetes in production best practices is a useful next read.
The connective thread is deployment โ models increasingly ship as containerized workloads, which loops straight back to Kubernetes.
Together: one backbone, four labels.
Pick a primary cloud based on your job or job market. Do Cloud Practitioner or AZ-900, add AI-900/AI Practitioner if AI is on your radar. Brush up on Linux, networking, and Git alongside it.
Go deep: SAA-C03 or AZ-104. Build 2โ3 real projects โ Terraform, a CI/CD pipeline, a containerized app. Start poking at EKS or AKS.
KCNA first if Kubernetes is new, then CKA. Learn Helm, GitOps, monitoring, basic security. Maps directly to DevOps Engineer Professional or AZ-400.
ML Engineer Associate or Generative AI Developer Professional (AWS), or AI-103 (Azure). Practice deploying models on Kubernetes, basic MLOps, RAG โ the stuff interviews actually ask about.
Add your second cloud's core cert (AZ-104/AZ-305 if you started AWS, or vice versa). Optional: Solutions Architect Professional, an Azure Expert cert, or CKS.
Five phases in sequence: not four certificates โ one connected skill set.

Learn AWS IAM, and Entra ID clicks almost instantly โ same model, different name. Same story with VPCs โ VNets โ Kubernetes Services.
One path, four vendor labels on the same ideas.

Same path, different exits.
Multi-cloud shops want engineers who move between AWS and Azure without a ramp-up. This roadmap signals exactly that โ plus Kubernetes fluency and AI readiness on top. If you're still weighing which cloud to prioritize, MicroDegree's comparison of AWS vs Azure vs GCP career growth in 2026 breaks down where the demand is headed.
Result: you're not "the AWS person" boxed into one lane. You're qualified for cloud engineer, DevOps, solutions architect, platform engineer, and increasingly AI/ML engineer roles โ the profile most senior positions actually want.
Four certifications, one road. Start with the foundations, go deep on one cloud, use Kubernetes as your bridge layer in AI, then extend into cloud number two.
Follow that order, and 24 months from now you won't just have four credentials โ you'll have one skill set employers are actually hiring for.
Yes โ the exams differ, but concepts like IAM, networking, and IaC overlap enough that one roadmap covers all four.
Whichever your employer uses or your job market demands. The second cloud always comes faster.
Yes, if cloud-native or DevOps roles interest you โ Kubernetes is the common runtime layer across both clouds.
Yes. AI Practitioner and AI-103 validate skills now showing up in cloud engineering job descriptions, not just data science ones.
It's better not to prepare for all four simultaneously. Start with one primary cloud, build your core cloud skills, then move to Kubernetes and AI. Once your fundamentals are strong, adding a second cloud becomes much easier because many concepts overlap.
Not sure where to start? Explore MicroDegree's Cloud, DevOps, and AI Agents paths to find the sequence that fits your background โ or drop a comment with your current role and target certification, and we'll help map out a personalized version of this roadmap.
Know someone stuck choosing between AWS, Azure, Kubernetes, and AI certs? Send this their way.