I train models and put them where the user already is.
One of them reads 171 hieroglyph signs at 98.2%a, inside a phone browser, with no photo leaving the device.
Sources
- the project's own README, checked 2026-08-28
Work
Ten projects, in three groups. Each group makes a claim you can check against every project underneath it.
Models that leave the notebook
Trained, then put somewhere a person can open.
-
Wadjet
98.2% reported hieroglyph classification accuracy. Source: the project's own README. Checked 2026-08-28.
-
Incident Forecasting Pipeline
0.8684 R² on the current dataset. Source: the public repo and the internal deliverable log, which agree. Checked 2026-08-28.
-
Cerebral Stroke Prediction
Systems that stop
Every one of these knows a condition under which it refuses to continue.
-
Grounded RAG PDF Q&A
17/17 golden eval, twelve grounded questions and five adversarial ones. Source: the project's own README. Checked 2026-08-28.
-
workflow-doc-agent
9 tests, none of which touch the network. Source: the project's own README. Checked 2026-08-28.
-
Multi-agent reporting pipeline (client work, anonymized)
11 pipeline stages. Source: the delivered build, read directly. Checked 2026-08-28.
Second opinions
Each one measures the same thing twice, because one measurement is not evidence.
-
Paired time-series synchrony analysis
7,991 paired computations, sixty-one channels across one hundred and thirty-one segments. Source: the delivered notebooks and their saved outputs. Checked 2026-08-28.
-
Sentiment Alignment, Amazon reviews against BERT
-
Student Success Analytics
-
T2D Adipose Tissue Research
About
I live in Desouq, a town in Kafr El Sheikh, up in the Nile Delta. Cairo time, GMT+2.
I am reading for a BSc in Artificial Intelligence at Kafr El Sheikh University. I started in September 2023 and expect to finish in July 2027.
Most of my work is the unglamorous half of machine learning. Cleaning data. Engineering features. Checking whether a number means what it looks like it means. Then getting the result somewhere a person can open it.
Wadjet is the clearest example. I trained three image classifiers, exported them to ONNX, and ran them client-side through ONNX Runtime Web. The photo never leaves the phone, and inference costs the server nothing, so it survives on a free tier. A model trained in a Delta town ends up running inside a browser somewhere else. That is the part of this I find worth doing.
I went through the DEPI Data Science track from 2024 to 2025, led the cohort team, and finished with Best Project and Best Member. The following summer I did two internships, one at the National Telecommunication Institute and one at Egypt's Ministry of Communications and Information Technology.
Since July 2025 I have been a Data Science intern at Zetta Global, forecasting operational incidents across 129a locations. Real data behaves differently from a downloaded dataset. It grows, the distribution shifts underneath the model, and the honest number goes down rather than up.
Since October 2025 I have been Head of Programming at IEEE Kafr El Sheikh, where I plan the curriculum and teach the Python sessions. Before that I ran the C++ track. I have competed in IEEEXtreme, the 24-hour global programming competition.
I freelance as well. Three completed contracts, two of them rated 5.0b.
What I am aiming at is production systems rather than research. I graduate in July 2027.
Sources
- the public incident-prediction repo, checked 2026-08-28
- Upwork profile, read directly, checked 2026-08-28
Now
- Building
- Extending the Zetta forecasting pipeline into driver risk scoring and segmentation, reading from and writing to a PostgreSQL warehouse. Also building this site.
- Learning
- Nothing new to report this month.
- Available for
- Data science and AI engineering roles, and freelance work.
Updated 2026-08-28.
Contact
Write to me at the address below. There is no form, because a form can break quietly and an address cannot.