Your DevOps Journey

The Phantom Obstacles

Everyone wants to talk about the tools. Kubernetes this, Terraform that. Nobody wants to talk about the psychological traps that derail actual DevOps careers. Here are four that trip up smart people.


“DevOps Is Only for Developers”

Wrong.

Coding helps, but it’s not the gate. Some of the best DevOps engineers I know came from tech support, NOC, or sysadmin backgrounds. Why? Because half this job is troubleshooting under pressure, and you don’t learn that from LeetCode.

The reality:

  • CI/CD pipelines are more YAML archaeology than software engineering
  • Cloud orchestration is about knowing what breaks, not writing elegant code
  • Monitoring and observability reward pattern recognition from hard-won experience

Rule: Breadth beats depth in the early years. Worry about specialization later.


“I Don’t Match the Job Description”

Job descriptions are written by committees playing bingo with buzzwords. “Kubernetes, Terraform, Python, Go, AWS, Azure, GCP, CI/CD, SRE mindset, 10 years experience” for a mid-level role.

Ignore the noise. Focus on:

  • What you have done: concrete problems solved
  • What you can learn: demonstrated curiosity matters more than current knowledge
  • Fit for the environment: a hungry engineer at a growing company beats a checked-out expert at a stagnant one

Rule: Apply to jobs where you meet 50% of the requirements. The other 50% is what makes the job interesting.


“If It’s Not FAANG, It’s Not Real”

Big tech is a career path, not the career path. The FAANG obsession misses something critical: smaller companies let you touch everything. No silos. No 6-month approval processes to provision a load balancer. You build, you break, you fix, you learn.

What you actually get at smaller shops:

  • Scope that would take years to earn at Big Tech
  • Direct line to decisions (for better or worse)
  • Visible impact: you can point to systems you built

Rule: Optimize for learning velocity, not brand prestige. Your next job cares what you can do, not where you learned it.


“Follow the Crowd”

The industry chases shiny objects. GitOps! Platform Engineering! Internal Developer Platforms! Every year there’s a new mandatory paradigm.

Here’s the truth nobody tweets: Your environment is unique. Legacy systems, compliance requirements, team dynamics, budget constraints, these dictate your tooling more than Gartner quadrants ever will.

Pragmatism > Trend-chasing:

  • Use what works for your constraints
  • Evolution beats revolution: continuous improvement compounds
  • The “wrong” tool you can operate beats the “right” tool you can’t

Rule: There is no one-size-fits-all DevOps. Anyone selling one is selling something.


The Bottom Line

DevOps careers stall when people let external noise override internal judgment. Develop broad skills, ignore gatekeeping job posts, pick environments where you grow, and solve the problems in front of you, not the problems Hacker News thinks you should have.

FAQ

Q: How do I transition from traditional operations to platform engineering?

A: Start by automating repetitive tasks with infrastructure as code. Learn one cloud provider deeply, master CI/CD fundamentals, and focus on building internal developer platforms that abstract complexity. The key is shifting from “keeping lights on” to “enabling developers to ship faster safely.”

Q: What platform engineering skills are most in-demand for 2026?

A: Kubernetes administration, Terraform/Pulumi for infrastructure automation, strong GitOps practices, and developer experience (DX) design. Companies also value platform engineers who understand observability, cost optimization, and can build self-service tooling that reduces cognitive load for development teams.

Q: Should I learn to code before starting a DevOps career?

A: Basic scripting (Bash/Python) is essential, but you don’t need software engineering-level coding skills to start. Focus on automation, troubleshooting, and infrastructure concepts first. Many successful DevOps engineers started in support or system administration roles without formal programming backgrounds.