Ahmed Abdelaziz

I translate between the people with the problem and the code that fixes it.

Pharma commercial operations by day, Computer Science and Business by degree. I find the grunt work eating your team's days, then pick the right fix: a rule, an off-the-shelf tool, or a model I train myself.

In progress

Part of my day job is checking debit notes against compliance rules: the same checks on every note, every week. Grunt work. I'm automating it at work with the tools my company approves, and that build is still in progress. Here I'm rebuilding it in public, on made-up data so nothing from work leaves work, and without those limits, so I can test the real question: a rule, an off-the-shelf model, or one I train myself? The write-up compares all three, including where my model loses.

  • Compliance check rebuild on synthetic data: A rule, an off-the-shelf model, or one I train myself? Tested on synthetic debit notes, with metrics and failure cases in the write-up. · target 30 Nov 2026

Work you can check

Now

  • Commercial operations by day.
  • CS + Business at the University of Toronto.
  • CompTIA Security+ (SY0-701) · earned Oct 2025
  • Azure AI Apps and Agents Developer Associate (AI-103) · working toward it
  • Cisco Certified Network Associate (CCNA) · working toward it

How I handle data

Everything I build in public runs on synthetic data, and nothing from an employer goes into it. This site loads no trackers, ads or third-party scripts, and it sends a strict content security policy.