DEV Community: Cristian

Road to Golden Jacket

Cristian — Sun, 27 Sep 2026 19:03:32 +0000

More than four years after passing the AWS Cloud Practitioner, learning and earning AWS certifications along the way, I finally achieved the Golden Jacket!

I've been earning a few certifications each year, and this year I looked at the ones I had left and thought: Why not get them all? And, of course, get a cool/weird jacket along the way!

Now, I want to share some thoughts about each one based on my personal experience, what you need to know for each certification, and also which ones I found the toughest!

Foundational Certifications

Cloud Practitioner

This is the entry level certification, basically you just need to know the main AWS services, what they are used for, and some basic cloud concepts, security, pricing, and the AWS Shared Responsibility Model.

You don't need years of experience, but if it's your first certification, you may want to prepare well because of the nerves of facing a new challenge!

AI Practitioner

Also an entry level certification, but focused on AI, Machine Learning, and Generative AI concepts in AWS.

You will need to understand general AI concepts like LLMs, prompt engineering, RAG, model evaluation, and also AI security and governance.

If you already did the Cloud Practitioner, you are fine with the basic AWS services, but you will also need to add some AI services like SageMaker, Bedrock, and Amazon Q.

Associate Certifications

Solutions Architect Associate

I faced this one just a few months after the Practitioner, and at that time it was a challenge for me.

Probably the best general-purpose technical AWS certification. It focuses on designing secure, resilient, high-performing, and cost-optimized architectures.

It covers a lot of services, and it's not just about knowing what they do individually, but also understanding how they connect and work together.

It's not the easiest Associate certification, but it's the one that will give you a great understanding of the AWS ecosystem.

Developer Associate

This was my next step just after the Solutions Architect Associate, and it felt a lot easier.

You need to focus on fewer services. Lambda, DynamoDB, and API Gateway are the main ones you need to know about: how they work, their limits, and also how they interact with other services.

If you already have the Solutions Architect Associate, this is an easy one to get next!

CloudOps Engineer Associate

I first tried the old version (SOA-C02) and then recently the new one (SOA-C03), and I didn't find a significant difference. Just some updates to services that have changed over the years, but the main focus is still the same.

You'll be tested on operating, monitoring, maintaining, and troubleshooting workloads running on AWS.

It's not a hard one, especially if you work on these kinds of tasks in your day-to-day job.

Data Engineer Associate

I truly recommend doing this one before the ML and Generative AI Professional certifications, because for this one you need to learn and understand how to ingest and prepare data for different use cases.

You need to understand services like S3, Glue, Athena, Redshift, Kinesis, and EMR, and especially how to move, transform, store, and analyze data using them.

It also covers data security, governance, monitoring, and choosing the right services depending on the use case.

So why especially before ML? Because you need that data to train ML models, and you'll have a better understanding of this part when preparing for the ML certification.

Machine Learning Engineer Associate

Now that you already know how to prepare your data, here you have to learn how to train models and the different options you have for that.

For this, you need to know all about SageMaker, which is the main service for training models in AWS and the main focus of this exam.

The exam itself is not a hard one (as it's still an Associate), but this used to be covered by a Specialty certification, and it still feels a bit like that. You need to know about other services, yes, but the main focus is on SageMaker and its different options, as well as general ML concepts like hyperparameter tuning, model deployment, and model monitoring.

And now it's time for the hard ones:

Professional Certifications

Solutions Architect Professional

The advanced architecture certification. I didn't feel that what you need to know is much more than in the Solutions Architect Associate, but you have to know everything very clearly.

The difficulty here lies in the long and tricky questions this certification has. And there are 75 of them, so you probably won't have much time left at the end. In my case, I didn't even have time to review the questions.

You need to know about a lot of AWS services and how to connect them across multiple Regions, accounts, and VPCs, while keeping the solution scalable, secure, and cost-effective.

DevOps Engineer Professional

Also a Professional exam, so the questions are long and tricky, but I consider this one easier than the Solutions Architect Professional, probably because it covers fewer services, and also because, in my case, I work with Platform Engineering.

You mainly need to know about automation with CodeBuild, CodeDeploy, and CodePipeline, and also how to manage your infrastructure with CloudFormation.

But I have to say that there are other tools in the market that are much more widely used than the ones AWS offers, with a special mention here to GitHub and Terraform. So this is a good certification, but maybe in this particular field there are better certifications outside AWS.

Generative AI Developer Professional

The advanced Generative AI certification, focused on designing, implementing, and operating production-ready GenAI solutions on AWS.

Unlike traditional Machine Learning certifications, the focus is not on training models from scratch, but on integrating foundation models into real applications and business workflows.

The main service here is Amazon Bedrock. You need to know about Knowledge Bases, Agents, Guardrails, and how to implement all these concepts with Bedrock.

You also need to understand concepts like RAG, prompt engineering, embeddings, vector databases, model evaluation, and AI security and governance.

Specialty Certifications

Security Specialty

Even this one is an Specialty certification, I particulary found this one not really dificult, probably because was one of the last ones I did, and all the previous ones always includes questions about security.
Also because with time I gained experience in this field.

Here you need to deep dive into topics like identity, infrastructure protection, encryption, logging, monitoring, and incident response.

And about services you need to learn about IAM, KMS, GuardDuty, Security Hub, WAF, Shield, CloudTrail, and AWS Config, and understand how to use them together to secure AWS environments.

Advanced Networking Specialty

I'm still including this certification because I obtained it as part of my Golden Jacket journey, and it's still available today. But soon we will have to say farewell to it, as it is planned to be retired.

That doesn't mean networking isn't a relevant field. In my opinion, this is one of the most complicated exams, and it covers really relevant content that, even if you don't use it in your day-to-day work, is important to know.

This certification covers complex AWS and hybrid network architectures. You need to learn deeply about VPCs, Transit Gateway, Direct Connect, VPNs, and Cloud WAN, and also understand when each one is the best solution for for each case.

I faced this certification last because I had heard it was the most difficult one. And that could be true, but as I took it three years after the Solutions Architect Professional, I actually found this one "easier" because I had gained a lot of experience and knowledge during that time.

The Road to the Golden Jacket

When I started with AWS certifications, the Golden Jacket was never the goal (I didn't even know it existed). It was just about gaining knowledge and improving my skills.

It wasn't until last year that I learned about it and thought: "Hey, I'm not that far from having them all!" So I decided to go for it and pass all of them.

The main goal should always be to learn and understand the topics. Also, if you are working with AWS, getting certifications over the years can make a lot of sense, and as you gain experience, they become easier to prepare for.

Some of you may also ask: How do you get the jacket once you have all the certifications? At the moment, there is no public page where you can simply request it.

As far as I know, you can ask AWS about it if you work for an AWS Partner, if you are an AWS Community Builder, or you may have the opportunity to get one at re:Invent Las Vegas.

So, as a final recommendation, I would say that you should first go for the certifications that best suit your expertise and can have the biggest impact on your career. Whether you focus on architecture certifications or ML and AI certifications should depend on what you want to learn and where you want to grow, not just on getting a silly golden jacket.

Long is the way we have come!

AWS vs Azure vs GCP: Professional Architect comparison

Cristian — Sun, 10 May 2026 16:06:21 +0000

After passing the 3 big Cloud Architect Professional certifications, I wanted to write about that experience, and also compare the 3 clouds and some of the differences they have.

All Professional exams recommend 2 years of experience with their cloud, but if you already know one of them, you already understand many concepts.
For example, EC2, VMs and GCE. You already know that all are IaaS compute and what they are for, you just need to learn that in AWS you could have more control over them, and in GCP they are simpler with less to care about.
It's not that one is worse or better, they are different models to create solutions.

Organizations, Tenants and Projects

The big difference I saw, fundamental to learn about, is how each cloud structures all their resources.

AWS is built around accounts, grouped under Organizations, where each account acts as a strong isolation boundary.
Azure, on the other hand, revolves around identity: everything lives under a tenant (Entra ID), with subscriptions used to organize resources and billing.
GCP takes a different approach, where projects are the core building block, sitting under an organization and optionally folders.

AWS isolates by accounts, Azure organizes by identity, and GCP structures everything around projects.

Questions and duration

Azure and GCP are both 120 min and around 50 questions (it depends randomly), and AWS is 180 min and 75 questions.

Here I have to say that the AWS questions are sometimes extra long, and even with one hour more you have less time to read all of them.

Also mention that the Azure questions are not only questions with a few answers to pick. There are also questions to drag and drop, order, say if a scenario is correct, and also a big case study with more details and a few questions related to it.

AWS is longer and heavier, Azure is more scenario-based, and GCP is more about picking the right architecture than knowing every detail.

Cost and Prerequisites

AWS: $300, but if you already passed another AWS exam you have a 50% voucher in your account, so check it and use it!
The proper path here would be to take the Practitioner, the Architect Associate and then the Architect Professional. With the 50% discount for each one you pass, the 3 certs would cost $325, so it's almost the same price for all of them.

If you play it smart with AWS discounts, going from zero to Architect Professional can cost almost the same as a single Professional exam.

Azure: The price for the exam is $165, BUT you would also need to pass the AZ-104 Administrator first, so it's mandatory to add another $165 for that cert and the total for both is $330.

GCP: $200 and no prerequisites, easy and simple. You can try the Cloud Engineer Associate first, which is related and could be an easier first approach to Google exams, but it's up to you, it is not a prerequisite.

Expiration and recertification

About the three certifications, they have different durations until you have to recertify them.
AWS: 3 years, and you have to take the full exam again.
GCP: 2 years, but you would recertify it with a shorter and cheaper exam.
Azure: 1 year, but it is recertified with a short and free exam!

So here we can also see 3 models: longer validity with a full exam to recertify, or shorter validity with an easier exam.

Quick summary

Cloud	Price (USD)	Duration	Questions	Validity	Recertification	Prerequisites
AWS	$300	180 min	75	3 years	Full exam again	None
Azure	$330*	120 min	~40–60	1 year	Free online renewal	AZ-104 required
GCP	$200	120 min	~50	2 years	Renewal exam	None

*AZ-104 exam costs $165 and is required to obtain the Azure Expert Architect certification.

All exams need a deep understanding of each cloud, many of their services and how they interact, and what would be the best solution for a specific problem.

In my opinion, the most challenging one is AWS, because it is the longest one, and the questions in some cases are too long for the time you have. At some point you are not only facing your knowledge and experience, you are also fighting against time and trying not to get tired of reading.

And with this comparison between the 3 certs, are you ready to take one of them?

I hope that now the pieces fit...