AI Engineer — Agentic Systems & Backend
I build agentic AI systems and the tooling that keeps them honest: MCP servers, agent skills, context-engineering pipelines, and verification gates that treat AI-generated code as falsifiable, not trusted.
Before this: 5+ years of production backend engineering (Node.js, Java) — microservices, Redis caching layers, API design and versioning — including 3.5 years fully remote for a Finland-based company, and leading a team that shipped a full ERP platform.
| Project | What it does |
|---|---|
| falsifiability-kernel | Executable gates, a falsifiability ledger, and blind cross-model review for AI-assisted development — a review process that actually terminates. |
| halal-graphify | Maintained fork of the Graphify CLI (codebase → queryable knowledge graph, used as grounding context for AI agents) — renamed "god nodes" to "hub nodes", auto-synced with upstream weekly, integrated a Godot/GDScript backend. |
| BroDashGraphify | Desktop dashboard + MCP server exposing those knowledge graphs to AI agents — with hard budget guards so an agent can never quietly spend your quota. |
| taste-engine | Zero-dependency web component that builds a preference profile through LLM-generated card battles. Works with local models. |
Also: a milestone-orchestration agent skill that decomposes projects into self-contained briefs, dispatches sub-agents, and verifies every result against git before moving on.
- 🎓 B.Eng. Computer Engineering — AASTMT, 2020 · MCIT Egypt Scholarship · 2× ACM ECPC contestant
- 📍 Alexandria, Egypt — open to remote & on-site
- 📫 LinkedIn · omarmhmd.om@gmail.com