14+ years building intelligent systems.
From algorithms to organizations.
LLMs & GenAI · Computer Vision · NLP · MLOps · Robotics
I came from applied mathematics and robotics - software is the craft I practice every day, but mathematics is where the thinking starts. I design systems where a wrong algorithm doesn't just produce a bad prediction, it can crash a physical machine. That background shapes how I think about AI: mathematically grounded, production-ready, and built for the real world.
Over 14+ years, I have moved from solving differential equations and programming robotic controllers to directing AI departments and managing R&D portfolios exceeding $12M. Today I lead teams delivering production-grade AI across LLMs, GenAI, computer vision, NLP, and predictive analytics - for governments, enterprises, and startups across MENA, USA, and Brazil.
Currently Manager of AI & Data Science at iHorizons, Qatar.
End-to-end model development, from research to production deployment.
RAG pipelines, multi-modal systems, open-source and commercial LLMs at scale.
From raw data pipelines to executive-level insights and visualizations.
Production-grade AI systems, observable and built to last.
Multi-cloud AI deployment and managed ML services across major providers.
The foundation - rigorous mathematical thinking applied to real-world systems.
Leading the AI & Data Science practice at iHorizons, a premier MENA IT solutions provider - delivering enterprise AI for government and private sector clients. Architecting and deploying multi-agent AI systems with autonomous planning, tool-use orchestration, and memory management. Leading LLM fine-tuning initiatives including SFT, LoRA, QLoRA, and RLHF-based alignment - optimizing open-source models for domain-specific tasks and deploying on-premise for enterprise clients. Overseeing RAG pipelines, multi-modal NLP systems, and large-scale data processing on GCP and Azure. Presenting AI strategy and project outcomes to C-suite executives and government stakeholders.
Recruited to build the Data Science & AI department from zero within Saudi Arabia's Vision 2030 ecosystem - handling hiring, technical standards, delivery frameworks, and best practices. Led a government-facing visual analytics project detecting urban visual pollution at city scale, with a modular pipeline built for extensibility without retraining. Led computer vision model training pipelines including YOLO, ResNet, and EfficientNet architectures. Designed Arabic NLP solutions for sentiment analysis, document classification, and conversational AI. Architected scalable ML platform infrastructure enabling rapid experimentation and production deployment.
Led the Data Science division delivering mathematically rigorous ML solutions across manufacturing, energy, and financial services. Built and deployed a predictive maintenance IoT system for a major MENA manufacturer - end-to-end from sensor data to executive dashboards, aligned with Industry 4.0 standards. Trained deep neural networks for time series forecasting and anomaly detection. Developed a financial advisory AI system ingesting 11 data sources, applying ML and quantitative models for prediction, risk analysis, and live portfolio management.
Directed robotics engineering, creating higher-level system designs and implementing AI models for autonomous robotics applications.
Led the AI, robotics, and automation division - delivering intelligent systems for industrial and commercial clients at the intersection of mathematical modeling and physical-world AI deployment. Created innovative system designs integrating computer vision and intelligent path planning. Applied machine learning to predictive maintenance, commercializing research into production industrial systems. Contributor to the Smart Automated Facade Cleaning System (US Patent 62830472).
An intelligent robotic system for autonomous high-rise building maintenance. (US Patent 62830472)
1 journal, 1 book chapter, 3 conference papers on robot intelligence and machine analysis algorithms.
International Best Researcher Award - ISSN, for outstanding contribution to research in robotics and machine intelligence.
Selected most meaningful from a broader portfolio: Google Professional ML Engineer, 5x Dataiku Data Science, 4x Nvidia AI including DGX AI Compute Systems and CNN Visual Recognition.
Persistent memory infrastructure for AI agents. MCP-native, local-first, hybrid BM25 + vector search. Zero-setup mode with SQLite - no API key or GPU required.
Continuous LoRA fine-tuning as a local service. Watches source files for changes, retrains on change, and hot-swaps adapters into a live OpenAI-compatible inference endpoint.
CLI tool that generates and optionally executes shell commands from natural language input. Practical, fast, and model-agnostic.
Path Optimization and Multi-Level Path Planning for The Steam Field Navigation Algorithm
Tool Wear Monitoring in Milling Machines Using Frequency Domain Vibrations Analysis
Robotic System for High-Rise Building Facade Cleaning
A Framework for Robotic Path Planning Based on Enhanced Fluid Potential Dynamical Models
Computational Time Reduction in Stream Field Navigation Through Dynamic Path Search and Selection in Kinematic Environments
"Leadership is a responsibility - not a position. A responsibility to deliver quality and to drive the growth of the people you lead."
I have directed an R&D budget of 12+ million USD across multiple organizations and continents. I lead by building trust first, planning with intention, and staying close enough to the work to be genuinely useful to my teams.
Whether it's a leadership opportunity, a research collaboration, or a complex AI problem worth solving - I am interested in conversations that matter.