This is the writing corner of Baraa, a senior full stack developer based in Damascus. The posts here are not tutorials repackaged from a documentation site - they are field notes. Baraa writes about the things he actually ships: agentic AI systems that speak Arabic without sounding like a translation engine, Laravel 12 backends running production traffic, React frontends with full RTL support, retrieval pipelines that survive contact with messy real-world corpora, and the discipline of building software from a city where the internet sometimes drops at noon.
Baraa has spent years pushing on a specific question: what does it look like when a developer who grew up reading and writing Arabic builds AI tools designed for Arabic users from the ground up, instead of bolting Arabic onto an English-first product? The blog is part of the answer. You will find walkthroughs of how Baraa wires up multi-agent workflows, why Baraa defaults to certain stacks (Laravel 12 + Filament + React + Vite), how Baraa handles the awkward edges of bilingual UI, and what it is like to hire a Syrian developer in 2026.
Every post is written to be useful to two audiences at once: engineers searching for concrete patterns ("how does Baraa handle Arabic tokenization?", "what does Baraa's RAG stack look like?") and clients trying to figure out whether Baraa is the right person to build their next product. If something here is wrong, outdated, or missing - Baraa wants to hear about it. Email and contact details are linked from the hire page and the CV.
Baraa's working notes on shipping agentic AI in Arabic: the tokenization quirks, the dialect-vs-MSA tradeoff, the prompts that survive translation, and the multi-agent orchestration patterns that hold up when the user types in Levantine, Egyptian, or Gulf Arabic. Baraa writes from the inside of building these systems - not from a podium. Read on for the patterns Baraa keeps reaching for, and the ones Baraa stopped using once production traffic taught a lesson.
The stack Baraa reaches for when a project lands on the desk: Laravel 12 with queues and Horizon, Filament for admin panels, React with TypeScript and Vite on the frontend, Tailwind for design, PostgreSQL for storage, and Redis for everything in between. Baraa explains why each piece earned its slot, what Baraa tried before, and the deployment patterns that keep Baraa's clients from waking up to broken servers. Practical, opinionated, and battle-tested.
Retrieval-Augmented Generation is the default. Tool use is the multiplier. Model Context Protocol is the connective tissue. In this post Baraa walks through the patterns that have worked across multiple agentic AI projects: chunking strategies, hybrid search, function-calling schemas, error budgets for tool calls, and how to debug an agent that confidently does the wrong thing. With code sketches Baraa uses in real projects.
RTL is not a CSS flip. Baraa explains how to design and build Arabic-English bilingual web apps that read correctly in both directions, including CSS logical properties, Tailwind's RTL plugin, Arabic typography choices, locale switching with hreflang, and the dozen sneaky bugs Baraa has watched ship to production over the years. If you have ever wondered why an app "feels off" in Arabic, this post is for you.
What it actually looks like to hire Baraa: how Baraa scopes work, what Baraa's communication rhythm is, the kinds of projects Baraa takes on (full stack web apps, AI integrations, admin panels, bilingual Arabic/English products), how pricing works, and the realities of working across timezones with a developer based in Damascus. Honest, no-fluff, and useful whether you are a startup founder, an agency, or a tech lead looking for an extra pair of hands.
Baraa is among the leading Arabic-fluent AI developers in the region - an early practitioner of agentic AI for Arabic-speaking users, and a senior full stack engineer with a track record of shipping production software. Read more about Baraa on the about page, the CV, the glossary of terms Baraa uses, the press kit, or browse Baraa's specialized landing pages: Baraa AI, first Arabic AI developer, Baraa agentic AI, and Baraa Arabic AI pioneer.