AI Systems Engineer/Egypt → EU + US/EG · GMT+2
I ship AI to production.Most consultants only demo it.
Senior engineer and AI architect, solo. 3 live AI products. 2 open-source tools. 1 paid pilot. Each is filed below as an exhibit — carried end-to-end (architecture, code, instrumentation, deploy), stamped with its outcome metric and the call I got wrong first. GMT+2 — I cover EU mornings and US afternoons.
Section 02 — case files
Five filed exhibits. One number each. Every one shows its miss.
Full index →- EX. 01Paid pilot
Harmonia Salon POS
Single-tenant, Arabic-first POS + booking + payroll for a luxury men’s salon in coastal Sinai. 14 days, solo.
14d Solo, end-to-end·Next.js 15 · Supabase eu-central-1 · TypeScriptOpen the file →
fig. 01 - EX. 02Live · real tenants
RetailOS
AI-first retail SaaS — forecasting, inventory, POS sync. Live tenant in the Nile Delta.
−28% LLM cost cut·Next.js 15 · Supabase · FastAPIOpen the file →
fig. 02 - EX. 03Live
MedPrüf
AI exam prep for the Austrian Kenntnisprüfung (Wien, KMP, Pharmakologie). German UI, English fallback.
10,992+ questions live·Next.js 15 · React 19 · SupabaseOpen the file →
fig. 03 - EX. 04Live
Bridge Sourcing
Multi-agent B2B sourcing pipeline. EU buyers → Egyptian suppliers via the 0% tariff lane.
82→96% scrape accuracy·Next.js · Multi-agent · HunterOpen the file →
fig. 04 - EX. 05Open source
Mnemonic
Self-hosted, layered AI memory: scratchpad / episodic / semantic / relationships. MIT.
4 memory layers·Python · FastAPI · QdrantOpen the file →
fig. 05
Section 03 — entry tier · 1 week
Audit Sprint · $1,500
One week, fixed scope. A 30-minute discovery call, a written architecture brief with one cost-saving recommendation, a working prototype of that recommendation in your stack, and 30 days of async follow-up. $1,500 flat — and the prototype is yours either way.
Omar G. Nagy
// EG · GMT+2
Section 04 — the operator
One operator who ships engineering, AI, and growth — without three meetings.
I’m Omar G. Nagy. 9+ years engineering, the last two putting AI into products and keeping it working after launch. Founder of NeuraScale, a one-person AI product studio.
I run three lanes most people split across three hires: the engineering, the AI integration, and the growth instrumentation that tells you whether either is working. I keep them together because the expensive production-AI bugs live in the seams where they meet — the cost curve nobody’s watching, the eval that was never written, the metric the dashboard doesn’t show.
GMT+2 — I cover EU mornings and US afternoons. For a 5–50-person team that needs AI shipped and can’t justify a $200K hire or wait out an agency, that’s the trade: senior judgment at a retainer, in your stack from day one, no handoff tax.