Al Samawal (Sam) Abdalla

Who I am and what I am aiming for

I am a Linux and infrastructure person based in Vienna. I founded Ataraxium, a project to deploy Linux systems for AI agents with infrastructure as code, and I am now building it properly on Terraform.

Target role: Linux engineer, DevOps engineer or platform engineer where Terraform is part of the daily work.

Certification: I am preparing for the HashiCorp Certified: Terraform Associate (004). I do not hold it yet.

Ich lebe in Wien und bereite mich auf die Prüfung HashiCorp Terraform Associate (004) vor. Ziel: eine Stelle als Linux-, DevOps- oder Platform-Engineer mit Terraform.

Background

• Founder of Ataraxium, registered as a business name in Ireland (Companies Registration Office no. 767020, July 2025). Built from Cork (2025) and Dublin (April to August 2026).
• Ran my own company in Belgium, wound down in December 2025.
• Diploma in Computer Hardware Engineering, South Kent College, United Kingdom (1997).

Learning path, 10 to 18 November 2026

The spine of my study is one course, followed in its own order:
HashiCorp Certified Terraform Associate (004) Full Course by sthithapragna (free, 7 h 05 min). The official 004 review list is my checklist.

Every day follows the same loop: watch the chapters, answer the questions out loud in English and German, build it in a practice lab, break it on purpose, fix it, repeat it on my real Linux server, and write a short daily report.

Labs and proof points

Only real work is marked done. Everything else is shown with its planned date.

Done

• done Level 1 lab, practice setup (10 Nov 2026). A small Terraform project using two official providers that need no cloud account (hashicorp/local, hashicorp/random). It creates a random node id and writes a status file. The exercise starts broken on purpose (no pinned providers, a wrong resource reference, a missing required variable); a check script runs init, validate, plan, apply, destroy and apply again before it passes. Tested on Terraform 1.12.2, the version the exam uses.
• done Study plan for 10 to 18 November, built on the course chapters above, with English and German question sets for each lab level.

terraform {
  required_version = ">= 1.5.0, < 2.0.0"
  required_providers {
    local  = { source = "hashicorp/local",  version = "~> 2.5" }
    random = { source = "hashicorp/random", version = "~> 3.6" }
  }
}

resource "random_id" "node" {
  byte_length = 4
}

resource "local_file" "status" {
  filename = var.output_filename
  content  = "Ataraxium lab: ${var.lab_name}\nnode_id: ${random_id.node.hex}\nstatus: ready\n"
}

Level 1, fixed version (shortened).

In progress

• in progress Daily job-style reports, one per day from 10 Nov: tasks, what was done, problems and fixes, next step.
• in progress Level 1 on the real server: first look at the server (Linux version, disk, memory) on 10 Nov; Terraform installed in its own folder on 11 Nov.

Planned, with dates

• 11–17 Nov Lab levels 2 to 8, as in the learning path.
• 15 Nov Small monitoring piece: a timed health check that tests every few minutes that the server answers and writes the result to a log, with a short README and a screenshot.
• 15 Nov Lessons learned, week 1: real problems from the daily reports, each with cause, fix and what I would do next time.
• 17 Nov Destroy-and-rebuild video, 3 to 6 minutes, narrated: the code, plan, destroy, the system gone, apply, the system back and working. No secrets on screen.
• 18 Nov Final full destroy and rebuild on the server.

The Ataraxium project

What it is: a business-to-business tool that creates infrastructure as code to deploy Linux systems for AI agents. Terraform is its engine; it does not compete with Terraform.

Honest status: early stage. It is being rebuilt from the ground up, one lab level at a time, following the course above. What exists today is lab level 1. The rest is the plan.

What it has to prove (targets, not yet done)

1. Terraform code that creates a Linux server for an AI agent: provider, variables, modules, outputs.
2. That server's setup written as code, so every deployment comes out identical.
3. An automatic pipeline: tests run on every change, and it deploys only when the tests pass.
4. State handled safely, with a recorded destroy and rebuild.
5. Security basics: firewall, safe remote login, secrets never in the code.

How the lab levels build it

• Level 1 · first Linux server configuration: providers, variables, plan and apply
• Level 2 · input checks on the server specification
• Level 3 · API keys for the AI agent kept out of code
• Level 4 · a reusable "AI agent server" module
• Level 5 · remote state and locking
• Level 6 · import, drift and rebuild
• Level 7 · one workspace per customer or environment
• Level 8 · full mock exam and final written report

Contact

• Email: al.samawal.sam.abdalla@gmail.com
• LinkedIn: linkedin.com/in/al-samawal-abdalla-a0a6a3156
• Website: ataraxium.ai
• Location: Vienna, Austria

Kontakt gerne auf Deutsch oder Englisch.