About me
Patrick Saade
AI engineer. I build AI products end to end, and lead AI work inside teams part-time.
ProductArc is my practice: one engineer, end to end.
My story
From Paris to the Gulf, one production system at a time
I started in Paris, with an MSc in AI Systems at EPITA and an internship at Quantmetry (Capgemini Invent) building MLOps infrastructure. In Luxembourg, at Augment, I built my first text-to-SQL chatbot and RAG assistant.
At GlobeMed in Beirut there was no AI team. I pitched a working translation model to the CEO, secured funding in six weeks and co-founded the AI/ML department.
In 2025 I went independent, and GlobeMed kept me on as a contractor. On a UAE Federal Ministry's transformation program I led the AI modules, including a chat-with-database system over national workforce data. I also built products of my own, like Parlo and SYD, and worked with Zaka.ai on AI products and training.
When a team already has something in progress, I also take consulting and advisory work: architecture reviews, evaluation setups and AI roadmaps.
Values
What I care about
Plain language
Every decision explained without jargon, so you always know why something is built the way it is.
Measured, not guessed
I test on real questions and real data before anyone relies on the system.
Documented
Clean code and clear docs, so your developers can take over from day one.
Privacy by design
Sensitive data stays where it belongs: on the ministry's forecasting platform, the AI only saw aggregated statistics.
Track record
Where I've worked
Contract work, and the teams I built AI with before that.
Senior AI Consultant (Contract) · GlobeMed Group
- Lead AI initiatives in Bahrain and Saudi Arabia: intelligent document parsing and claims case summaries.
- Built the LLM evaluation framework; structured the department's codebase, architecture and standards.
AI Engineer · Zaka.ai / MadeBy
- Shipped customer-facing LLM features in a commercial SaaS (React, FastAPI, AWS, OpenAI), used by hundreds.
- Designed a 100-participant Certified AI Consultant curriculum; mentored 20 interns and capstone students.
AI Engineer (Contract) · Inheaden, for a UAE Federal Ministry
- Architected and owned a text-to-SQL chatbot over the national workforce database (7 million individuals): 80 to 90% accuracy on 1,000+ evaluated queries, in English and Arabic, with milestones presented to the ministry's leadership.
- Rebuilt a failed forecasting platform in two weeks, about 70% faster, keeping row-level data out of the LLM.
Machine Learning Engineer · GlobeMed Group
- Co-founded the AI/ML department: pitched a working translation model to the CEO; full funding secured in six weeks.
- Fine-tuned MarianMT for Arabic-English medical translation; grew training data from 1K to 15K records.
AI Engineer · Augment Digital Facilities Services
- Built a LangGraph RAG assistant over equipment manuals and a text-to-SQL chatbot for a BIM platform.
Machine Learning Engineer, Intern · Quantmetry (Capgemini Invent)
- Built the internal MLOps platform on AWS and Kubernetes (Terraform, Ansible, Kubeflow, Seldon Core).
Tech stack
Tools I work with
End-to-end delivery across the stack, from model training and evaluation to APIs, full-stack apps, infrastructure, and production deployment.
AI & ML
Frameworks
Web & App
Data & Infra
Have an idea that deserves to get past the demo?
Book a 30-minute call and tell me what you want the AI to do.