I build things because sitting with a question is never enough. First class honors graduate, King Abdulaziz University, and drawn to the problems that matter most and get the least attention.
How I got here.
It started with a robot. I built a computer vision pipeline that let a machine navigate and complete tasks with no human input. That got me asking harder questions about everything underneath it.
That curiosity pulled me toward deep learning, then toward language, then toward Arabic, the language most models still treat as secondary. I fine-tuned models, published research, and built multi-agent systems that reason over real data. Each project was not a requirement. It was a question I needed to answer.
Saudi Arabia is at a rare moment. The infrastructure, the ambition, and the investment are all converging. I want to be one of the people building the systems that make that moment last. Not demos. Not experiments. AI that organizations can actually rely on.
A question I cannot stop thinking about.
Where I applied it in the real world.
At SABIC, one of the world's largest petrochemical companies, I worked within the Center of Excellence for Sales, an enterprise-scale environment with strict standards for data quality, governance, and reliability.
I led the design and development of an AI-driven customer analytics solution to predict promoter likelihood from over 4,000 survey responses. The solution combined a machine learning model with Power BI dashboards and achieved 80% accuracy, eliminating a fully manual analysis process. AI had not been applied to this workflow before I joined.
I also collaborated on a CRM reporting governance framework, establishing access controls, data quality standards, and authentication mechanisms, and supported the testing phase of SABIC's SAP S/4HANA enterprise rollout (the STAR project).
What that curiosity produced.
An intelligent platform for foreign investors entering Saudi Arabia, simplifying market understanding, investment evaluation, opportunity discovery, competitor analysis, and regulatory navigation through a single agentic experience. Multiple specialized agents handle each part of the journey. Validated at 85.7 to 100% pass rates across all agents using G-Eval evaluation.
A multi-agent courtroom where AI lawyers argue real cases through structured legal reasoning. A Judge agent orchestrates seven deterministic trial stages; Plaintiff and Defense agents build arguments, manage evidence, and track credibility scores. Every ruling follows consistent admissibility logic. Full audit trail rendered in real time, a demonstration of AI reasoning in high-stakes, structured decision environments.
An autonomous agent that identifies vulnerabilities in web applications and knows exactly where to stop. OWASP knowledge baked into a RAG layer enables intelligent security analysis. A hard scope validation boundary fires before every action. Full audit trail. The agent acts freely, not dangerously.
A computer vision pipeline for a robot that navigated and completed tasks with no human input during competition. First place at the Saudi Games. The first proof that autonomous systems can be trusted to act in the real world.
Collaborative research and published work.
Programs completed along the way.
A few words from people I've had the pleasure of working with.
Open to the right opportunity.