📍 United Arab Emirates · 🧭 Technical Lead @ WA.Technology · 🧠 AI Platforms & High-Scale Backend Engineering

I build production software where AI, backend architecture, data pipelines, and developer tooling meet. My background spans long-running PHP platforms, distributed/event-driven systems, streaming infrastructure, and modern agentic/LLM workflows.

I focus on systems that need to be reliable under load, maintainable by teams, and practical to operate in production.

• AI platforms & agentic systems — orchestration, model routing, RAG/evidence workflows, local inference, and AI-assisted developer tooling.
• Backend & distributed systems — Laravel/PHP, Node.js/TypeScript, Python services, queues, Kafka, caching, and high-throughput event pipelines.
• Search, verification & data systems — hybrid retrieval, claim/evidence workflows, PostgreSQL, OpenSearch, embeddings, and ingestion pipelines.
• Infrastructure & delivery — Docker, Cloudflare, AWS, CI/CD, observability, deployment safety, and production troubleshooting.

Founder / Technical Lead · AI-powered fact-checking · News analysis · Arabic & English

Labeeb helps people and organizations understand news and online information more clearly, distinguish reliable information from misleading content, and move from claims to source-backed evidence.

Built around the realities of Arabic and English media, Labeeb combines fact-checking and news intelligence with an evidence-first approach designed to make analysis easier to understand and trust. The goal is not simply to label information, but to give users clearer context around what is being reported, what can be verified, and what evidence supports the conclusion.

Explore Labeeb: Website · About · How Labeeb Works · AI Ethics & Transparency · LinkedIn · Building Labeeb

Core development is maintained privately under the labeeb-io organization.

PowerShell · OpenVINO · OpenVINO Model Server · Local LLM inference

A Windows 11 toolkit for running INT4 LLMs on Intel Arc GPUs through OpenVINO Model Server, including installation automation, model switching, diagnostics, performance profiles, and OpenAI-compatible integration for IDEs and coding agents.

What it demonstrates: local AI infrastructure, GPU/runtime integration, automation, compatibility layers, and developer experience.

Go · Agent routing · Model gateways · CEL rules

A native Bifrost PreRequestHook that classifies agent work by role and capability, combines it with complexity routing, and selects model/fallback lanes dynamically.

What it demonstrates: agent architecture, model routing, gateway internals, deterministic classification, deployment safety, and Go plugin integration.

TypeScript · Next.js · React · Cloudflare · SQLite/D1

An open-source workspace for discovering, organizing, transforming, and sharing AI prompts, with versioning, search, prompt transformation, PWA support, and local/cloud deployment paths.

What it demonstrates: modern full-stack TypeScript, product UX, edge deployment, local-first workflows, and AI provider integration.

I use open-source contributions as a way to work across unfamiliar codebases, languages, and architectural styles — not just as a contribution counter.

• gemini-web2api — contributed merged features improving Gemini API interoperability and temporary-chat workflows in a JavaScript/OpenAI-compatible project.
• delegate-skills — contributed a Bifrost delegation skill for planning, architecture advice, and independent code review workflows.
• MoMoA-Researcher — explored/refactored authentication, LLM integration, and Firebase-related handling in a multi-agent research codebase.

These contributions complement my main project work by exposing me to different conventions, runtime models, review processes, and ecosystems.

A few posts where I document the engineering decisions, experiments, and systems behind the work — not just the finished result.

Building Labeeb: A Story of Human Vision and AI-Powered Development Product architecture · AI-assisted engineering · verification systems	RAG-Powered PHP Development PHP · RAG · developer workflows · AI-assisted software engineering
Turning Intel Arc into a Local LLM Runtime OpenVINO · Intel Arc · OVMS · local inference · developer tooling	Joining Web Summit Lisbon 2026 Web Summit · Lisbon · developer community · AI & software engineering

View all posts on LinkedIn →

Founder & Lead Developer · Android Open Source Project · 2013–2015

AOSB, originally known as ProBAM, was one of the projects that shaped me most as an engineer. It was a custom Android firmware project built across the Android/CyanogenMod ecosystem — not a single application. My work reached into the Android framework, ROM customization, OTA/update code, build and release tooling, and multi-device support across the project's open-source codebase.

That experience taught me how to work inside a large upstream platform, understand software below the application layer, handle compatibility across different devices, and build with an active user and contributor community around the project.

The project was publicly distributed through SourceForge, and its community support campaign recorded $1,960 raised from 75 contributions.

Explore the history: GitHub Organization · SourceForge Archive · Archived ProBAM.net · Community Support Campaign

My GitHub history goes back to 2011, so the profile also includes older projects that reflect the technologies and problems I worked on at the time.

• 🔐 ncryptd — PHP application encoder/obfuscation tooling; one of my longer-running public PHP projects.
• 🎬 video-js-swf — custom Flash player integration for Video.js.
• 🎬 flashls — HLS/Flash streaming work from an earlier stage of my media engineering career.
• 🎬 videojs-hls-levels — HLS quality-level selection tooling for Video.js.
• 📦 async-job-processing-with-batch-queuing — asynchronous job processing and batching concepts for high-throughput backend workloads.

• Prefer clear architecture and operational simplicity over unnecessary abstraction.
• Design for failure, rollback, observability, and maintainability, not only the happy path.
• Use AI aggressively as an engineering multiplier, while keeping testing, review, and final technical judgment in the software delivery loop.
• Comfortable moving between legacy systems and modern stacks when the business requires both.

• LinkedIn: linkedin.com/in/hanyalsamman
• Linktree: linktr.ee/hany.alsamman
• Website: codexc.com
• Email: hany.alsamman@gmail.com
• Calendly: 30-minute meeting