About
Three jobs, one thread: learn it fast, ship it properly
I am Ibrahim Maher Al-Bander, a social media marketer, accounts and inventory officer, and freelance AI automation developer based in Erbil, Iraq. I currently hold three roles at once at Help Tech Co. Ltd., and run an independent n8n automation practice on the side. None of that was the plan when I started. It is just what happens when you say yes to the next unfamiliar problem often enough.
Where it started
I studied Computer Engineering Technology at Northern Technical University in Mosul, graduating ranked 9th in my class with a capstone project building a CNN-based traffic-sign recognition system in Python. That project is the reason I was never intimidated by AI models later on; I had already spent months elbow-deep in how they actually learn, long before ChatGPT made that fashionable.
After graduating, I did what a lot of engineering graduates in Iraq do: I took the work that was in front of me. That meant freelance software testing on Upwork, building Telegram bots and small automations for fun, and eventually a marketing internship at YourVisio Media that turned into a full-time offer at Help Tech Co. Ltd.
How three jobs became one job
I joined Help Tech as an Assistant Sales Manager, handling procurement and vendor negotiation. A few months in, I picked up social media marketing because nobody else on the team wanted to touch it, and grew the company Facebook page more than 500 percent. Then the accounts and inventory role opened up with no system in place, so I built the company's QuickBooks bookkeeping process from scratch, with no formal accounting background, because someone had to and I was already the person who figured things out.
Somewhere in the middle of that, I started using n8n to stop doing the same reporting work by hand every week. That habit turned into a genuine skill, then into freelance client work, then into a small, real practice building production automations with OpenAI, Claude, and Gemini behind them. I am not trying to collect job titles. Every one of these roles showed up because something in the business needed doing and I was willing to learn it under deadline instead of waiting for someone more qualified to show up.
Ownership of unfamiliar problems
I have never waited for training before taking on a role. QuickBooks, n8n, prompt engineering, they were all things I learned on the job, under deadline, because the business needed them solved now.
Documentation over heroics
A system only I understand is not a finished system. Every automation I ship comes with a README and a one-click import, and every process I set up gets handed off in a way colleagues can actually run without me.
Numbers and words, together
Bookkeeping made my marketing reporting sharper. Marketing made my automations more useful to actual humans, not just technically correct. I do not see accounting, marketing, and automation as separate skillsets; they inform each other constantly.
Why AI tools are central to how I work now
I use ChatGPT, Claude, and Gemini daily, not as novelty but as working tools with specific jobs: content drafting and campaign analysis in marketing, structured data extraction in automation, and increasingly, direct software development. The clearest proof point is DAD LINK, a live e-commerce store I designed and shipped on Odoo with Claude Code doing most of the build work: product catalog, category pages, on-page SEO, and blog content, not a demo, a real store with real distributors.
I think about AI work in four layers: prompt engineering, context engineering, harness engineering, and loop engineering. Most people who say they "know AI" have only touched the first one. I write more about that distinction, and about the automation and marketing work itself, on the blog.