Short answer: most career changers get there in 6 to 12 months of consistent study. Learn the 15-25 AWS services that carry real architectures, get the Solutions Architect Associate certification, build one system you can explain end to end, and learn Infrastructure as Code.
That is the whole path. The rest of this guide is the detail — and the honest parts most roadmaps leave out.
What an AWS cloud engineer actually does
Before the roadmap, understand the job. Most people picture it wrong.
A cloud engineer is not someone who memorised AWS services. Day to day, you are:
- Designing infrastructure that meets requirements — handles the traffic, stays available, does not overspend
- Building it with code, not by clicking through the console
- Securing it — controlling what can reach what, and what each component is permitted to do
- Monitoring it and responding when something breaks
- Reducing cost, because someone is always looking at the bill
Notice what is missing: nowhere does anyone ask you to recite what all 200+ AWS services do.
Free assessment
Find the AWS domains holding you back.
Get a focused study plan based on your real weak spots — not another generic course checklist.
Step 1: Build the fundamentals (4-8 weeks)
Skipping this is the most common reason people stall. AWS sits on top of fundamentals — if those are missing, everything above feels like memorisation.
Networking. IP addresses, subnets, DNS, ports, firewalls, public vs private. AWS networking is these concepts wearing AWS names. If subnets confuse you outside AWS, VPCs will confuse you inside it.
Linux. Most cloud workloads run on Linux. You need to be comfortable at the command line — navigating, permissions, processes, logs, SSH.
One scripting language. Python is the usual choice. You are not building applications; you are automating tasks and reading other people's code.
If you already work in IT, development or support, much of this transfers and you can move faster. If you are starting from zero, budget real time here. It pays back.
Step 2: Learn the services that actually matter (6-10 weeks)
Here is where most people lose months.
AWS has over 200 services. Trying to learn all of them leaves you more overwhelmed and confused than when you started. In the real world, roughly 15 to 25 services handle the vast majority of what companies actually build.
The ones worth your time first:
| Category | Services |
|---|---|
| Networking | Route 53, VPC, subnets, Security Groups, ALB |
| Content delivery | S3, CloudFront |
| Compute | EC2, Auto Scaling, Lambda, (later ECS/EKS) |
| Databases | RDS, DynamoDB |
| Integration | SQS, SNS, SES |
| Security | IAM, Security Groups, Secrets Manager |
| Operations | CloudWatch, CloudFormation or Terraform |
That is the list. Learn those properly and you can read most production architectures.
Learn them by problem, not by definition
This is the part that changes outcomes.
Do not learn "S3 is object storage." Learn "static files and dynamic content have completely different cost and performance profiles, so we separate them — static goes to S3, dynamic stays with the application servers, and we do not want servers burning capacity serving banner images."
Do not learn "Auto Scaling adds instances." Learn "a load balancer distributes traffic across whatever servers exist, but it cannot create servers — so without Auto Scaling, launch day takes the site down."
Every service exists because without it, something breaks. Learn what breaks.
The fastest way to internalise this is to study one complete system rather than a service list. I walked through exactly that in AWS Architecture Explained: How a Real Ticketing Platform Uses 15 Core Services — one customer followed from typing the URL to receiving their confirmation email, with every service justified along the way. If you can explain that system, you can explain most systems.
The two decisions interviewers probe most are covered in depth separately: EC2 vs Lambda and RDS vs DynamoDB.
Step 3: Get certified (4-8 weeks)
Certifications are not the goal. They are the thing that gets your CV read.
Hiring got harder in the AI era. Screening is noisier and portfolios are easier to fake, which makes verifiable proof that you passed a supervised exam a genuine signal.
If you are completely new: start with AWS Certified Cloud Practitioner ($100). It is foundational, it builds confidence, and passing earns a 50% discount voucher toward your next exam — so it partly pays for the one that follows.
The one that matters: AWS Certified Solutions Architect – Associate (SAA-C03) ($150). This is the highest-value single certification for cloud engineering roles. It tests exactly what you learned in step 2 — which service solves which problem, and why you would choose it over the alternative.
Skip Professional-level certifications for now. They assume hands-on production experience and are best attempted after a year or two in the role.
How to actually study for it
Not by watching a 40-hour course start to finish. Most people buy one, get overwhelmed within a week, and stall.
The efficient approach is gap-based: find out what you already know, then spend your time only on what you do not. Answer practice questions, review why the wrong answers were wrong, and re-test on your weak domains. Active recall beats passive watching by a wide margin, and it cuts study time dramatically.
Step 4: Build one complete system (4-6 weeks)
One realistic system you can explain end to end beats five tutorial projects you followed along with.
A strong portfolio project touches:
- DNS and content delivery
- A VPC with public and private subnets
- A load balancer with Auto Scaling behind it
- At least one database, ideally two of different types
- Something asynchronous — a queue or event-driven function
- IAM roles with least-privilege permissions
- Monitoring and alarms
The ticketing platform above is a good template because it forces every one of those decisions. But build something you find interesting — you will be talking about it in interviews, and genuine interest shows.
The thing that matters is not the diagram. It is being able to justify every choice. Expect to be asked: why EC2 and not Lambda here? Why two databases? What happens if that Availability Zone fails? What stops two people buying the same seat?
If you can answer those, you are ahead of most applicants — including some with the job title already.
Step 5: Learn Infrastructure as Code (2-4 weeks)
This is the step that separates people who get hired from people who stay stuck at "certified but no offers."
In a real job, none of this gets built by clicking through the console.
Say you spent days configuring an architecture by hand. It works. Three months later someone asks for an identical copy for testing. You cannot remember every setting. You miss something. The test environment does not match production, so when something breaks there it tells you nothing useful.
Infrastructure as Code fixes that. You write a file describing everything — every service, its configuration, its connections, its permissions — and hand it to Terraform or CloudFormation, which builds exactly what you described. Need another environment? Same file. And because it is a file, every change is tracked: who changed what, when, and why.
Rebuild your step 4 project in Terraform. That single exercise moves you from "did a course" to "worked how teams actually work."
Step 6: Apply with proof
Most applications lead with courses completed. That is the weakest possible opening.
Lead with the system. "I built a ticketing platform on AWS that handles traffic spikes with Auto Scaling behind an ALB, uses RDS for the event catalogue and DynamoDB for seat availability, and prevents double-selling with conditional writes. The whole thing is defined in Terraform." Then link the repo and the diagram.
That is a different conversation from "I have the Solutions Architect Associate."
Do not ignore adjacent routes in. Cloud support, DevOps-adjacent roles, internal transfers, and help desk positions at cloud-heavy companies all get you working with real infrastructure. Many people reach cloud engineer through one of these rather than landing it directly.
Realistic timeline
| Phase | Time | What you are doing |
|---|---|---|
| Fundamentals | 4-8 weeks | Networking, Linux, scripting |
| Core services | 6-10 weeks | The 15-25 services that matter |
| Certification | 4-8 weeks | Cloud Practitioner, then SAA-C03 |
| Portfolio system | 4-6 weeks | Build one thing end to end |
| Infrastructure as Code | 2-4 weeks | Rebuild it in Terraform |
| Total | 5-9 months | Studying consistently alongside a job |
Some people move faster because fundamentals already transfer. Some take longer because life happens. The variable that actually predicts success is consistency, not intensity. Ninety focused minutes a day beats an eight-hour weekend binge followed by two idle weeks.
The honest parts
A certification alone will not get you hired. It gets you past screening. What converts is being able to talk through something you built.
You will feel underqualified longer than you expect. Everyone does. The gap between "I passed the exam" and "I feel like an engineer" closes through building, not through more courses.
The market rewards proof over credentials. Which is good news if you are self-taught, because proof is something you can manufacture deliberately.
AI has not reduced demand — it moved it. AI workloads run on cloud infrastructure. Someone has to architect, secure and operate that. The bar has risen on demonstrating you can actually do the work, which is exactly what steps 4 and 5 are for.
Where to start today
If you are starting from zero: fundamentals first, then the core services, then Cloud Practitioner.
If you already work in IT or development: skip ahead to the core services, then go straight for SAA-C03.
Either way, the single highest-leverage thing you can do this week is stop trying to learn all of AWS and start understanding one complete system. Read through the ticketing platform walkthrough, and for every service ask yourself: what breaks without this?
Answer that consistently and the exam gets easier, the interviews get easier, and the job stops feeling out of reach.