AI & Data Engineer from Cairo.
I like taking work that feels annoyingly repetitive and turning it into a system that just runs โ so humans can spend time on the parts a computer can't do yet.
๐จ teaching CS ๐ฆ wiring up AI automation workflows ๐ฉ exploring data-intelligence tooling ๐ช turning experiments into products
The stuff I keep coming back to:
|
AI Pharaoh Actually teaching a machine to read hieroglyphs. |
Meridian Wings Optimizing a network when the future is uncertain. |
AutoEDA Give it a dataframe. It asks the annoying questions first. |
|
DataPrepToolkit Messy data in. Usable data out. |
Healthcare ML API Models wrapped in an API that actually deploys. |
Movie Search Semantic movie search. Meaning instead of keywords. |
๐๏ธ AutoBI lives on the portfolio rather than the public repo shelf. It's one of my main data-intelligence projects โ explore the full case study โ
I got tired of rewriting the same data workflow for every project. So I started building the pieces I kept redoing:
Each tool came from the question the previous one left unanswered.
- DataPrepToolkit โ prep, validate, profile (171 tests, MIT)
- AutoEDA โ explore, visualize, report (265 tests, MIT)
- AutoBI โ dashboards that survive contact with real data
Build ยท Think ยท Train ยท Ship ยท Automate
A small snapshot of what accumulates when I spend too much time inside repositories.
Repo languages, not a skill ranking.
The cards live in this repo and refresh daily via a GitHub Action โ no third-party stats service. Jupyter Notebook is excluded from the language chart.
Not everything needs to become a product. Sometimes I build something just because the problem looks fun.
- ๐ฏ trained a face recognizer that runs live โ face-recognition-streamlit
- ๐ก sniffed network packets and turned them into ML features โ Network-Packet-Analyzer
- ๐ broke my own encrypted connection to understand MITM โ secure-network-communication
- ๐ฎ built an ASCII platformer because... why not โ Console-Platformer-Quest
Things on my mind lately:
- โ making AI agents actually useful in messy, real-world workflows
- โ automation that survives the ugly edge cases
- โ turning notebooks into products people can rely on
- โ going from a dataframe to a decision people trust
