GitHub Agentic PR Dataset
~1.96M GitHub pull requests from AI coding agents (Claude Code, Cursor, Copilot, Devin) and humans — joined to 6.7M commits and 55M file-level diffs. Open data (CC-BY-4.0) now used for agentic-AI and MSR research.
The research has a build side. These are the data products, tools, and applications I’ve shipped — each tracked with its real state, the way you’d read a pull request.
~1.96M GitHub pull requests from AI coding agents (Claude Code, Cursor, Copilot, Devin) and humans — joined to 6.7M commits and 55M file-level diffs. Open data (CC-BY-4.0) now used for agentic-AI and MSR research.
A stdlib-only Python job, run by a daily GitHub Action, that pulls my latest Medium articles and YouTube videos into versioned JSON and auto-commits — so the site stays current with zero manual upkeep.
Hands-on Android Studio builds that wire AI coding agents and on-device AI into real app workflows — published as walkthroughs for students and developers.
As Head of the Computer Center Department at the University College of Applied Sciences (UCAS), Gaza, I led the team that delivered the platforms below. This is team work — credit belongs to the engineers who built them; my role was setting direction, the architecture and IT-policy calls, and getting them shipped.
Head of Computer Center · UCAS Gaza
Directed the migration of a legacy Oracle system to a modern web platform serving 500+ staff — spanning infrastructure, IT policy, and rollout.
One central login across every university system — with SMS and authenticator-app 2FA, access and session logs, and role-based permissions backed by a full audit trail.
Real-time monitoring of payments and revenue from the eSadad gateway — a transaction ledger, analytics, financial-period comparisons, and raw-data auditing for the finance team.
Automated Google Classroom synchronization producing course, instructor, and student performance reports — turning e-learning activity into administrative insight.
Google Drive–backed document management with direct scanning, in-place Google Docs editing, and time-boxed fine-grained sharing — digitizing the paper archive at zero storage cost.
An internal link shortener with real-time analytics, QR-code generation, multi-format exports, and GitHub-based automated deployment.
Hardened and modernized the infrastructure — firewalls and port hardening, AWS failover servers, Cloudflare Pro / WAF, Zabbix + ManageEngine monitoring, Tailscale secure access, Oracle 19c, and automated daily backups.
More to come — presentations & talks, and notes, are next. In the meantime, the dataset and publications go deeper.