Available for new projects

AI Product Engineering

I build the AI layer of your product.

From idea to launch, end to end: the assistants, agents and machine-learning features, the backend and data behind them, and the deployment that keeps them running. For startups and product teams that want AI shipped, not just demoed.

Book a 30-min call Try my live products

ProductArc is my AI engineering practice: one engineer, end to end.

Patrick Saade · The engineer behind ProductArc.

LinkedIn GitHub Download CV More about me →

Dubai · Paris · BeirutEnglish · French · ArabicRemote · Clients worldwide

Built AI systems for a UAE Federal Ministry and GlobeMed ($3bn in claims a year), and my own products used by real people.

The problem

Most AI projects stop at the demo.

Turning an idea into a demo is quick. Then real users arrive, and the cracks show.

Where most projects stop

1IdeaWhat the AI should do
2DemoImpresses the roomWhere most projects stop
3ProductUsed every day, end to end

Answers you can't trust

The assistant sounds confident and is sometimes wrong, and nobody can say how often.

Costs and delays that grow

What felt fast and cheap for ten testers isn't for a thousand users.

No way to measure it

Without evaluation, every prompt change or model swap is a guess.

I take it the whole way, end to end.

Live now

Don't take my word for it. Try it.

Live

Parlo

Build a set of questions, share a link, and people answer by chatting with a friendly AI instead of filling out a form.

LLMsFull-StackReactFastAPI

Try it

Live

SYD

A preventive-health assistant that answers only from CDC and USPSTF guidance, cites every source, and safety-checks every answer.

RAGGuardrailsHealthcareLLM Evaluation

Try it

Video

TableTalk

Ask a database questions in plain English while AI agents write and check the SQL.

Agentic AIMulti-AgentRAGText-to-SQL

Watch

Work

Client work

Client · contract

National workforce data, in plain language

UAE Federal Ministry

The problem: Analysts needed answers from the national workforce database, covering 7 million people, and every question meant writing SQL by hand.
What I built: I architected a text-to-SQL assistant in English and Arabic on a federal transformation program, and rebuilt a failed forecasting platform in two weeks, about 70% faster, with row-level data kept out of the LLM.
The result: 80 to 90% accuracy on more than 1,000 evaluated questions, presented to the ministry's leadership.

Text-to-SQLRAGAgentsPrivacy by design

Screenshots not shown under NDA.

Client · employee, then contractor

Claims and medical translation AI, in production

GlobeMed, processing nearly $3bn in claims a year across 12 countries

The problem: Claims had to be handled faster and medical documents translated accurately between Arabic and English, with no AI team in place.
What I built: I built intelligent document parsing and claims case summaries now in production in Bahrain and Saudi Arabia, plus the LLM evaluation framework behind them, and fine-tuned a MarianMT Arabic-English medical translation model in PyTorch on training data I grew from 1K to 15K records.
The result: Running in two new markets on shared engineering standards.

Started by pitching a working model to the CEO, funded in six weeks, and grew into GlobeMed's AI/ML department.

Document AILLM evaluationFine-tuningHealthcare

Screenshots not shown under NDA.

More projects

Process

How I get your AI past the demo

Each step exists to fix one of those problems: answers you can't trust, costs that grow, and no way to measure.

Scoping

What it should do, what data it needs, and how you'll know it works, agreed before anything is built.

Architecture

Designed for accuracy, cost and speed at real scale, not just for ten testers.

Build

The AI and everything around it, inside your product, with something working to try every week.

Evaluate

Tested on real questions and real data, so you know how often it's right before anyone relies on it.

Handover

Deployed, documented and yours to keep, with the evaluation in place for every future change.

What you can count on

Scope agreed in writing before I write a line of code
Something working to try every week, not a status report
You own 100% of the code
Documented so your developers can maintain it
NDA signed on request

Services

What I build

The AI features your users interact with, built to hold up in production, and the software product around them.

AI assistants & chat-with-your-data

Let your users or your team ask questions in plain language over your documents and databases, and get answers they can check.

Full product build

AI is only useful inside a real product. I build the app, API, database, auth and deployment around it, so you get one working product.

AI agents

AI that doesn't stop at answering: it reads, decides and uses your tools and APIs, with guardrails around every step.

Document AI

Extract, classify, anonymize and translate claims, contracts and forms automatically, including scanned documents.

Custom & fine-tuned models

When off-the-shelf models aren't accurate enough for your domain, I train and evaluate one that is.

AI consulting & advisory

Already have something in progress? I review architecture, set up evaluation, and give your team a clear plan. Available as a short engagement or part-time.

Engagements

Ways to work together

Start here

Typically 6 to 10 weeks

MVP build

From an idea to a launched product with real users.

Typically 4 to 8 weeks

AI feature build

One clearly defined AI feature, from idea to production, with the scope agreed before any code is written.

Typically 1 to 2 weeks

Prototype rescue

Your prototype stalled. I measure where it fails, fix what matters most and leave you with a clear plan for the rest.

Ongoing, 2 to 3 days a week

Embedded AI lead

I join your team part-time to build and lead your AI work: architecture, evaluation, standards, delivery.

Discuss your project

Recommendations

What people I've worked with say

“Patrick demonstrated strong technical expertise, ownership, and problem-solving abilities. He was capable of taking projects from requirements and architecture through implementation and deployment, while collaborating effectively with both technical and business stakeholders.”

Salil Trehan

Principal AI Engineer · Inheaden

“One of the key attributes that set Patrick apart was his diligent approach to documentation. He ensured that every piece of code and every process was appropriately documented, making it easier for other team members to understand and build upon his work.”

Jade Ghadry

Co-Founder & CTO · Augment

“He has shown his ability to work independently and as a team member. He successfully delivered on time with high-quality results.”

Aubin Mazet

Lead ML Engineer · Quantmetry (Capgemini Invent)

Full recommendation letters are available on request. Ask me for them.

FAQ

Before you book

Where are you based? Do you work with clients abroad?

I'm based between Dubai, Paris and Beirut, and I work remotely with clients anywhere in the world. I schedule calls in your time zone and keep the work moving asynchronously in between.

Do you sign NDAs?

Yes, before you share anything sensitive.

Who owns the code?

You do, 100%, from day one.

Can you work with our developers?

Yes. Most of my work happens inside existing teams, and I document everything so your developers can take it over.

What do you need from us to start?

A 30-minute call, a description of what you want the AI to do, and some example data if you have it.

I'm not technical. Can you handle everything?

Yes. I build the AI, the app around it, the backend and the deployment, and I explain every decision in plain terms.

Can you start from just an idea?

Yes. Scoping is the first step: what it does, what data it needs, and how you'll know it works.

What happens after launch?

You get the code, documentation and evaluation setup. I hand over to your team or stay on part-time.

Do you also consult?

Yes. Architecture reviews, evaluation setups and AI roadmaps, as a short engagement or part-time inside your team.

Contact

Have an idea that deserves to get past the demo?

You'll talk to me directly, from the first call to handover.

Book a 30-minute call to talk it through, or send me a message and I'll get back to you.

Book a 30-min call

contact@productarc.io

Dubai · Paris · Beirut · Remote