Why DevOps?

When I first got into development, I didn't know what to specialize in yet. I worked with C, system programming, and algorithms during the 42 core curriculum. I liked understanding how things work underneath. Not interfaces, not design — mechanisms, infrastructure, what keeps a service running.

That's what DevOps is. The link between code and production. Deploying, automating, monitoring, securing. It's not a support role — it's a role that impacts the entire chain.

What drew me in wasn't just the technical side. It's also the position. A DevOps/SRE/Platform Engineer touches everything: code, infrastructure, CI/CD, cloud, security. Long-term, it leads to roles like cloud architect or platform lead — positions where you make structural decisions.

But it's not just a career calculation. I genuinely enjoy what I do. Configuring a reverse proxy, writing a pipeline, debugging a container that won't start — that's the kind of problem I like solving.

Why Not Something Else?

AI — I like the tool, I use it daily, and there's a lot to build with it. But the sector doesn't suit me. Too much hype, massive investments for companies that still aren't profitable — OpenAI lost $8 billion in 2025 despite $13 billion in revenue, with projected losses through 2029. And technically, it's heavy on algorithms and mathematics — not my strength, not what drives me.

Cybersecurity — I worked with assembly at 42, did a few security-related projects. But the core of the job — investigating, hunting for vulnerabilities, analyzing traces — isn't what gets me up in the morning.

Full-stack — That's what I'm aiming for as a secondary skill, in a T-shaped approach. But after infrastructure, not before. Understand what's under the hood first, build interfaces later.

Learning by Layers

I didn't follow a roadmap I found online. I moved forward through dependency logic: each step builds a solid foundation for the next.

This isn't theoretical. If you don't understand how a Docker container works, you can't use Compose properly. If you don't know what a root user is, you can't configure Traefik or secure a container with a non-root user. Each layer builds on the previous one.

Linux / Bash — The foundation. You can't do DevOps without being comfortable in a terminal. 42 gave me that — not complete mastery, but enough to get by, read logs, write scripts, navigate a system.

Docker / Docker Compose — The first tool that changes everything. Understanding containers, images, networks, volumes. The Inception project at 42 laid the groundwork, Glasck gave me real-world practice.

AWS, Terraform, Ansible — Cloud and Infrastructure as Code. Provisioning AWS resources with Terraform, configuring machines with Ansible. A single project that had me working with all three.

Monitoring — Prometheus, Grafana, Loki. Set up on Glasck, not from day one as I should have, but the experience showed me why it's essential as early as possible.

Each step was driven by a project or a technology to explore concretely. Not a theoretical curriculum.

It's not a checklist. It's a logic: each step prepares the next.

Where I Stand — March 2026

This section is intentionally dated. The rest of the article describes a mindset that won't change. My skill level and priorities, however, evolve.

  • Done: 42 core curriculum completed — C, C++, basic system administration, networking fundamentals. Docker/Compose in production on Glasck. An AWS project with Terraform and Ansible. Prometheus/Grafana/Loki monitoring in place.
  • In progress: preparing for AWS Cloud Practitioner
  • Next steps: Kubernetes, CI/CD (GitHub Actions, GitLab CI), GitOps (ArgoCD/Flux)

Being Honest About Your Level, and Prioritizing

I'm not an expert in any of this. Linux, I know the basics to get by. Docker, I can write a Compose file and debug, but I haven't touched everything. Terraform, Ansible, AWS — I have one project behind me, not ten.

I learn by doing. Each project pushes me forward, each problem forces me to dig deeper. It's not linear, it's not fast, but it's concrete.

And everything can't be learned at once. You have to choose. There are technologies that interest me, but that I'm deliberately pushing back — not out of disinterest, but out of priority.

Multi-cloud — No GCP or Azure before mastering AWS. AWS is used by 85% of enterprises worldwide — it's the standard. One cloud well understood is worth more than three skimmed over.

Kubernetes — I'll get there, but Docker Compose is enough for my current needs. No point adding complexity without a use case.

Advanced certifications — One at a time. Cloud Practitioner first, Solutions Architect after.

Choices guided by the market, available time, and what has the most impact right now.

Long-Term Vision

DevOps isn't an end goal for me. It's a foundation. The long-term objective: evolve into a cloud architect or platform lead role. Positions where you no longer just execute — you structure, you decide, you shape the technical architecture of a company.

Getting there requires a solid technical foundation, an understanding of business stakes, the ability to think about user experience, and knowing how to step back and see the bigger picture. That's what I'm building right now.