IEEE ICONET26AI music genre classifier, presented at an IEEE conference
82% accuracy · presented at IEEE ICONET26
Work
14 projects from study, internships and freelance work. Each one has a page: the problem, what was built, the result.
IEEE ICONET26AI music genre classifier, presented at an IEEE conference
82% accuracy · presented at IEEE ICONET26
Final-year project · A+Edit the transcript, and the video edits itself
Final-year project · graded A+
In productionClients don't cancel. They just go quiet.
Live in production at Wakilni

End-to-end ML pipeline with personalised recommendations
R² 0.97 in 5-fold cross-validation

Upload a transcript PDF, get a multi-sheet Excel report
PDF in, multi-sheet Excel out

A webcam coach that tells you what to fix, out loud
Real-time pose scoring with spoken corrections
Raspberry Pi 5 · offlineA robot pet that understands Arabic and English
Runs fully offline, on the device

One medical history from many hospitals
Three hospital databases, one history

Public-sector digital transformation, in three languages
Home-screen features in three languages

Transactions, roles and reports for a restaurant floor
Every order is one atomic transaction

My own programming language, interpreter and web IDE
A full interpreter and a web IDE

A genetic algorithm that builds a clash-free week
Clash-free weekly timetables

Monitoring, users, backups and logs behind one Bash command
SHA-256-verified backups, one command

The router forwards everything and reads nothing
Encrypted messages hidden inside images