Every LinkedIn post about AI sounds the same. "Excited to leverage cutting-edge LLMs to transform your business." Cool. What does that even mean?
I'll tell you what I actually do.
AI Engineering — Not the Buzzword Version
I build AI systems that run in production. Not prototypes that break when you look at them wrong.
RAG Systems: Retrieval-Augmented Generation is where I spend most of my time. Real RAG — not just "stuff everything into a vector store and pray." I design chunking strategies, retrieval pipelines, re-ranking layers, and evaluation frameworks. The stuff that makes the difference between a chatbot that hallucinates and one that gives you the right answer every time.
I've shipped RAG systems across banking, humanitarian operations, and eCommerce. Each domain has different tolerance for hallucination. Banking: zero. Humanitarian: lives depend on it. eCommerce: your customer service team will revolt.
AI Agents: Not the "look, my agent can use a calculator" kind. Production agents with guardrails, fallback chains, tool-use protocols, and observability. Agents that know when they don't know something. Agents that ask for permission before doing something irreversible.
The gap between a demo agent and a production agent isn't technical. It's architectural.
Full-Stack Development — The Backbone
AI doesn't float in a vacuum. It needs infrastructure.
I build full-stack systems around AI — Laravel backends, Next.js frontends, RESTful and GraphQL APIs, database design, deployment pipelines. The boring stuff that makes AI actually work.
My stack: Laravel, Next.js, React Native, PostgreSQL, Redis, Docker, cloud deployments. 20+ years of building software taught me that the flashy layer is never the bottleneck. The plumbing is.
I've built platforms that handle thousands of concurrent users, real-time document processing, multi-tenant architectures, and systems that need to work in low-bandwidth environments (try doing humanitarian tech in a conflict zone — you learn fast).
Mobile Apps — React Native
Cross-platform apps that actually feel native. Not the "it works on both platforms" kind of works. The kind where users can't tell it's not native.
React Native, with proper native modules when needed. Offline-first architectures. Real-time sync. The stuff that makes an app useful instead of just functional.
The Difference
Here's what separates me from the average AI engineer or full-stack dev:
I've been doing this for 20 years. Since 2005. I've seen every hype cycle, every "this will change everything" technology, every framework that was supposed to replace everything before it. Most of them are gone.
I design for the worst case, not the demo. When you're building for banking or humanitarian operations, failure isn't a learning opportunity — it's a catastrophe. That mindset carries into every project.
I'm in the MENA region. I build for this market. I understand the infrastructure challenges, the regulatory landscape, the user behavior. I'm not importing Silicon Valley patterns that don't translate.
I don't oversell. If your project doesn't need AI, I'll tell you. If a simpler solution works better, I'll build that instead. My job isn't to sell you technology — it's to solve your problem.
What I'm Looking For
Complex problems. Systems that need to actually work. Projects where the stakes are real — whether that's money, lives, or just someone's professional reputation.
Not "can you build me a chatbot." I can. But you probably don't need one.
The projects that excite me: RAG systems that need to be accurate, agent architectures that need to be safe, platforms that need to scale, and mobile apps that need to work offline.
If that sounds like your project, let's talk.
What I build: AI systems (RAG, agents, LLMs), full-stack web platforms, mobile apps, APIs, databases, and the infrastructure that holds it all together.
Where I build it: Amman, Jordan. For the world.
What's the biggest gap you see between AI demos and production systems? I'd love to hear your war stories.
More like this on yabasha.dev — no fluff, no "excited to share," just real systems and real failures.