About · AI Systems Engineer · Egypt · GMT+2 · open · for hire
Oneoperator.
I’m an AI systems engineer. 9+ years building software, the last two spent putting LLMs into products people actually use, and keeping them there after the demo (the part most of the field is still bad at). I run NeuraScale solo from Egypt: I do the architecture, write the code, wire the instrumentation, and ship it. No three-meeting handoff between the person who decides and the person who builds, because they’re the same person.
What that buys a 5–50-person team: senior judgment at a retainer, not a $200K-a-year hire and a six-month ramp. GMT+2: I cover EU mornings and US afternoons. One engineer covers both halves of the workday: the London standup and the New York call, same day, normal hours. Egypt isn’t the discount; the time zone is the edge.
The work I point to is real, and I’ll tell you exactly what it is. 5 live products. 2 open-source tools. 1 paid pilot. The products are proof I can carry a system end-to-end (some run for family businesses, some are free builds traded for warm intros), not a customer logo wall I’m dressing up as revenue. Every number on the case files clicks through to the thing itself. That’s the whole pitch: receipts you can audit.
The dossier · photo on file · roles · tools
The operator,on file.
What I do
- Fractional CTO
- AI Systems Engineer
- Solo founder
- Multi-agent operator
What I reach for
- TypeScript
- Next.js 15
- React 19
- Supabase / Postgres
- Python · FastAPI
- Claude + OpenAI APIs
- Vercel edge
- Telegram bots
- Zoho mail · CRM
- PWA / RTL
- Plausible · GA4
- years engineering
- 9+
- products live
- 5
- open-source tools
- 2
- paid pilot
- 1
Career ledger · 9+ years · most recent first
Six eras,on record.
2024–present
era /06
Founder & AI Systems Engineer · NeuraScale
One-person AI product studio. 5 live products, 2 open-source tools, 1 paid pilot: architecture through deploy, all solo. A small agent fleet runs my own ops overnight, which is where I learned firsthand which agentic patterns survive production and which only demo.
Production AINext.js 15SupabaseMulti-agent
2022–2024
era /05
Independent AI & automation consultant · MENA + EU
Moved to LLM work the week ChatGPT shipped. ~12 clients across e-commerce, agencies, and ed-tech. The early ones taught me that the model is the easy 20%; the cost curve, the eval set, and the failure modes are the 80% that decides whether the feature is still running in month three.
LLM integrationOpenAIAutomationsEvals
2021–2022
era /04
E-commerce + digital transformation · Various, incl. Taager
Performance, scale, and retention work, including for Taager, a social-commerce platform that raised $6.4M. This is where I stopped optimizing for the benchmark and started optimizing for the number the operator actually watches.
PerformanceScaleRetention
2020–2021
era /03
Co-founder / e-commerce engineer · mBalloon
Ran a DTC brand end-to-end: Shopify, a custom backend, and the paid-acquisition stack on top. Owning the P&L taught me to read engineering decisions as money, the same lens I bring to an LLM bill now.
ShopifyDTCPaid acquisition
2019–2020
era /02
Front-end developer (freelance) · Multiple agencies
Migrated legacy jQuery and WordPress front-ends to React, the unglamorous work where you learn that shipping fast and shipping accessible are not a tradeoff if the foundation is right.
ReactMigrationsAccessibility
2017–2018
era /01
Founder · LibreFit
Computer-vision body measurement on OpenCV + Xamarin, my first time putting an ML model in front of real users. It did not work commercially, and that was the most useful failure I’ve had: I’d polished the algorithm and underbuilt the product. Ship the product, not the algorithm. I’ve led with that ever since.
OpenCVXamarinCV / ML
Education
BSc Computers & Informatics, Suez Canal University, 2014–2018, First Class Honours. The degree is on the record; the production work above is what I’d ask you to judge me on.
How the years stack · 2017–present
Three lanes.One pair of hands.
Every era on the ledger feeds at least two of these. That overlap, not any one skill, is what you’re hiring.
/01
Engineering
9+ years of it. React migrations, performance and accessibility work for agencies, then whole products: front-end to Postgres, architecture to deploy, solo. The floor everything else stands on.
/02
AI
On LLMs since the week ChatGPT shipped. ~12 client integrations before I built my own products on them; a first ML model in front of real users back in 2017. The through-line: models demo easily, and the cost curve, the eval set, and the failure modes decide what survives.
/03
Growth
I’ve run a DTC brand’s P&L, the paid-acquisition stack on top of it, and retention work for a platform that raised $6.4M. Engineering decisions are money decisions; I build for the number the operator watches.
The direct line · replies within a day
Put the operatorto work.
A team that needs senior hands without the hire, a business that needs its first real system, an AI feature that has to survive production. Same door.