Agents do the typing
I work with coding agents. They load my standards before they touch a project, a hook formats and lints every file they write, and they cite the rule code when they break one. I still review, commit, push and deploy myself.
About
I’m Aland. I work on platforms, pipelines and security in the Kurdistan Region, Iraq, and I keep going down a layer until the problem has a name. Then I write down what I found, so nobody has to dig there twice.
Perspective
The problems I like cross a boundary nobody owns. A deploy that fails because of DNS. A security control that is really a developer experience problem. A platform decision that removes a hundred future decisions, or quietly adds them.
I started in software engineering, then automated processes across finance, procurement, cybersecurity and operations at Asiacell, and moved into platform work at the Kurdistan Region Government’s DIT. The thread through all three: a platform is an operating model written down as tools, defaults and policies.
That is also why I write so much down. A fix in my head helps once. A fix written as a chart, a workflow or a rule helps every team after me.
Working method
When I don’t understand a system, I don’t stop at the docs. I build a small version, break it on purpose, and work down through the layers until the failure makes sense.
That is how PKI, Linux networking, Kubernetes internals, GitOps and supply-chain security stopped being theory for me. Inside the notebook
Rules I keep
I keep my engineering standards in one repository: 34 skills, every
rule with a code like K8S-006. These eight sit above all
of them.
Build the smallest thing that works. If I can’t say why an option exists, it goes.
Defaults work without extra switches, and the obvious name is the right one.
Turning safety off takes an explicit setting. Turning it on never does.
Pass native objects and plain CLIs through. No wrapper settings, no unmaintained wrapper actions.
Do the same thing the same way everywhere, so nobody has to remember which project does what.
When something behaves surprisingly, fix it. Don’t document the surprise.
Run it, render it, test it. Report what ran and what didn’t. Never claim untested behaviour works.
Open protocols, such as OCI, S3, Postgres, OIDC and OTLP, so nothing is locked in without a reason.
Habits
I work with coding agents. They load my standards before they touch a project, a hook formats and lints every file they write, and they cite the rule code when they break one. I still review, commit, push and deploy myself.
Some vendors geo-block the Kurdistan Region, so before I recommend a tool I check it works from here. It made me prefer open protocols and software I can run myself: OpenTofu over Terraform, OpenBao over Vault, any OCI registry over one vendor’s.
If I correct the same mistake twice, it becomes a rule or a check. A hard bug becomes a field note. The common path should be safe and boring; the exceptional path should still be understandable.
My platform approachFoundations & focus
Specimen 01 · Certification
A vendor-neutral certification from the Linux Foundation and the CNCF that verifies the skills to perform the duties of a Kubernetes administrator. It is performance-based: every task is solved live on a command line running Kubernetes, not picked from multiple choice.
Exam domains, by weight
The same ground as my platform work: cluster lifecycle, networking, and troubleshooting from the node up.
Specimen 02 · Certification
Validates the ability to design solutions on AWS based on the AWS Well-Architected Framework: architectures that are secure, resilient, high-performing, and cost-optimized, and the judgement to review existing solutions for improvements.
Exam domains, by weight
AWS sits in the infrastructure layer of my toolkit, next to Pulumi, Linux, and Cloudflare.
Specimen 03 · Degree
A Bachelor of Science in Software Engineering from the American University of Iraq, Sulaimani.
What it laid down
The foundation the rest of the stack is built on.