I build Agentic AI systems that don't just run—they close deals. With 20+ years of high-stakes business development in the GCC combined with cutting-edge AI engineering, I bridge the gap between technical possibility and business reality.
AI-Powered Mental Health Triage System
- Impact: A privacy-first, offline-capable triage agent implementing WHO mhGAP guidelines for low-resource regions.
- Tech: Next.js, WebLLM (Llama-3.2 running locally in-browser), Supabase, Client-side encryption.
- Why it matters: Demonstrates strict data privacy compliance and the ability to run sophisticated AI models on consumer hardware without server costs.
Production-Grade Agentic Workflow Platform
- Impact: Enterprise-ready dashboard for managing and observing AI agents, designed to reduce support ticket volume by 15-30%.
- Tech: Python, LangGraph, RAG with re-ranking, Docker, Role-Based Access Control (RBAC).
- Why it matters: Moves beyond "chatbots" to reliable, observable business processes. Includes a comprehensive evaluation harness to ensure agent reliability.
Modern Arabic AI Agent
- Impact: A high-performance conversational agent optimized for Modern Standard Arabic (MSA), tailored for regional GCC business needs.
- Tech: Python, Pydantic AI, Gradio, OpenAI (GPT-4o).
- Why it matters: Showcases rapid development of culturally aware AI solutions using the latest efficient frameworks.
- AI/LLM: LangChain, LangGraph, Pydantic AI, Flowise, n8n, Vapi.
- Backend: Python, FastAPI, Supabase, PostgreSQL (pgvector).
- Frontend: Next.js, TypeScript, Tailwind CSS.
- DevOps: Docker, Nginx, CI/CD.
- Project: Antigravity
"I don't just write code; I build systems that understand human intent."