AWS Certification Skill Tree

As a professional with more than 4 years of experience with AWS, I might attempt to get an AWS Certification sooner or later this year, so I'm evaluating all the possibilities out there.

AWS itself provides a useful guide to the recommended AWS Certification Paths, where all the top cloud job roles are listed together with a bried description of their responsibilities.

Even if such a resource is more than welcome, I find that the big picture is not as clear as it should be: I decided to help myself building a skill tree with the aid of the awesome diagrams.

After having manually scraped the internet to get AWS Certification icons¹, I started experimenting with diagrams and I finally came up with this result (click to enlarge).

Source code

from diagrams import Cluster, Diagram, Edge
from diagrams.custom import Custom

ADVANCED_NETWORKING = "advanced_networking"
CLOUD_PRACTITIONER = "cloud_practitioner"
DATA_ENGINEER = "data_engineer"
DEVELOPER = "developer"
DEVOPS_ENGINEER = "devops_engineer"
MACHINE_LEARNING = "machine_learning"
SECURITY = "security"
SOLUTIONS_ARCHITECT_ASSOCIATE = "solutions_architect_associate"
SOLUTIONS_ARCHITECT_PROFESSIONAL = "solutions_architect_professional"
SYSOPS_ADMIN = "sysops_admin"

SOURCE = "https://d1.awsstatic.com/training-and-certification/docs/AWS_certification_paths.pdf"


def build_node(name: str, label: str = "") -> Custom:
    return Custom(
        label=label,
        # This requires you to have an assets folder
        # with the icons named after the above literals
        icon_path=f"./assets/{name}.png",
    )


with Diagram(
    name=f"AWS Certification Paths\n{SOURCE}",
    filename="aws_certification_paths",
    direction="TB",
):
    DiveDeep = Edge(color="magenta", style="dashed")
    with Cluster("Optional for IT/STEM/Cloud professionals"):
        cloud_practitioner = build_node(CLOUD_PRACTITIONER)
    solutions_architect_associate = build_node(SOLUTIONS_ARCHITECT_ASSOCIATE)
    developer = build_node(DEVELOPER)
    sysops_admin = build_node(SYSOPS_ADMIN)
    cloud_practitioner >> solutions_architect_associate
    cloud_practitioner >> developer
    cloud_practitioner >> sysops_admin
    with Cluster("Architecture"):
        with Cluster("Solutions Architect"):
            (
                solutions_architect_associate
                >> build_node(SOLUTIONS_ARCHITECT_PROFESSIONAL)
                >> DiveDeep
                >> build_node(SECURITY)
            )
        with Cluster("Application Architect"):
            (
                solutions_architect_associate
                >> build_node(DEVELOPER)
                >> build_node(DEVOPS_ENGINEER)
                >> DiveDeep
                >> build_node(SOLUTIONS_ARCHITECT_PROFESSIONAL)
            )
    with Cluster("Data Analytics"):
        with Cluster("Cloud Data Engineer"):
            (
                solutions_architect_associate
                >> build_node(DATA_ENGINEER)
                >> DiveDeep
                >> build_node(SECURITY)
            )
    with Cluster("Development"):
        with Cluster("Software Development/Test Engineer"):
            (developer >> build_node(DEVOPS_ENGINEER))
    with Cluster("Operations"):
        with Cluster("Systems Administrator"):
            (sysops_admin >> DiveDeep >> build_node(DEVOPS_ENGINEER))
        with Cluster("Cloud Security Engineer"):
            (
                sysops_admin
                >> build_node(SECURITY)
                >> DiveDeep
                >> build_node(DEVOPS_ENGINEER)
                >> DiveDeep
                >> build_node(ADVANCED_NETWORKING)
            )
    with Cluster("DevOps"):
        with Cluster("Cloud DevOps Engineer"):
            (
                developer
                >> build_node(SYSOPS_ADMIN, label="(Optional)")
                >> DiveDeep
                >> build_node(DEVOPS_ENGINEER)
            )
        with Cluster("DevSecOps Engineer"):
            (sysops_admin >> build_node(DEVOPS_ENGINEER) >> build_node(SECURITY))
    with Cluster("Security"):
        with Cluster("Cloud Security Architect"):
            (
                solutions_architect_associate
                >> build_node(SECURITY)
                >> DiveDeep
                >> build_node(SOLUTIONS_ARCHITECT_PROFESSIONAL)
            )
    with Cluster("Networking"):
        with Cluster("Networking Engineer"):
            (
                solutions_architect_associate
                >> build_node(ADVANCED_NETWORKING)
                >> DiveDeep
                >> build_node(SECURITY)
            )
    with Cluster("AI/ML"):
        with Cluster("Machine Learning Engineer"):
            (
                solutions_architect_associate
                >> build_node(DEVELOPER)
                >> build_node(DATA_ENGINEER)
                >> build_node(MACHINE_LEARNING)
            )

I am mostly a visual learner: the following table shows some property I wanted to retain in the skill tree next to the visual element I chose for the encoding.

Property	Encoded with
Optionality of Cloud Practitioner certification	dedicated diagrams.Cluster
Job area	diagrams.Cluster
Job title	diagrams.Cluster
Dive Deep layer	styled diagrams.Edge

With this skill tree available, some considerations are now straightforward:

• after the optional Cloud Practitioner, you basically have three front doors: Solutions Architect, Developer or SysOps Administrator;
• you are expected to know how to architecture first, and only later start implementing;
• the so-called Security by Design approach is enforced also ensuring that many certified professionals have measured themselves with Security specialty;
• the Development, DevOps and Operations paths seem to be more general-purpose than the ones departing from the Solutions Architect front door;
• in this AWS vision, a Machine Learning Engineer is a specialized Cloud Data Engineer;
• these paths suggest that the people building and fine-tuning data solutions should try to stick to their role and to avoid wearing too many hats (e.g. DevOps certifications are not listed in their paths).

Finally, I tried to answer my original question: which certifications I should try pursuing first? To do so, it's useful reinterpreting the skill tree as a proper graph, e.g. by removing redundant edges and merging redundant nodes². After playing around with diagrams options, I highlighted the three paths that look closer to my experience so far (and thus represent the most probable choices I will invest time in): Cloud Data Engineer, Machine Learning Engineer and Application Architect.

Certifications to start with

Ignoring the retiring certifications, I think the first ones to pursue given my current role and interests should be chosen amongst Cloud Practitioner (to familiarize with exams process), and the "Associate trinity" made of Solutions Architect, Developer and Data Engineer.

The bottom line: there are many suggested ways to reach a given certification, you can either follow these recommended/recognized paths or build your own. Enjoy!

---

1. The entire collection is already available somewhere for sure, but I didn't search that much. ↩

2. This rendition make also clearer the four tiers of certification and their increasing difficulty: Foundational, Associate, Professional and Specialty. To further deepen their features and the differences in scope/depth assessed, you can refer to this useful post. ↩