About
Mehdi Nadifi
Enterprise AI Transformation & Technical Product Leader
I architect, launch, recover, and scale consumer and enterprise products across the Gulf, United States, and Europe. Over 18 years, I have contributed to 27 apps through 0-to-1 builds, platform expansion, product recovery, lifecycle ownership, and advisory roles, including three super-app initiatives.
My experience covers fintech, luxury retail, eCommerce, real estate, health, and enterprise AI. I have led teams of over 30, with up to 14 direct reports. I consistently clarify ambiguous mandates, define products, and deliver on fixed deadlines.
I began my career in Madrid, developing real-time trading platforms for Santander, Barclays, BBVA, and Société Générale. In New York, I built an AI investment personalization engine for JPMorgan Chase & Co., inside a mobile app serving 33 million active mobile users.
In the GCC, I led the development of an Arabic-dialect conversational super app from concept to production testing in six months, through a joint venture between Majid Al Futtaim and I.AM+, will.i.am's technology startup. I managed a cross-functional team of over 30, including 14 direct reports. At Chalhoub Group, I oversaw the product roadmap for seven luxury brands, generating over $60 million in incremental revenue. For Emarat Petroleum's EmCan program, I rebuilt the API foundation and restructured delivery to meet a fixed launch date, resulting in the app reaching number one in the UAE App Store and winning Best Loyalty Program in the Retail Fuel Sector at ReFuel Forum MENA 2024. Since August 2025, I have led Limitls AI, advising energy organizations in the UAE and KSA on enterprise AI, from opportunity selection to solution design, governance, and production deployment.
Arabic-first product development and GCC localization are core strengths. Building an Arabic-dialect conversational product in 2021, before ChatGPT's public launch, was my most significant test in this area.
Currently, I am building two products:
- Vigil Intel is a B2B geopolitical risk platform for institutional investors, covering 173 countries across ten risk dimensions. It uses an LLM Council architecture with independent LLM-as-judge verification to improve reliability and minimize weak outputs.
- Sona is a WhatsApp-native family AI operating system built as an independent working prototype, using specialized agents, persistent household memory, multi-model reasoning, and external tools. Development is paused.
I bring technical expertise and commercial insight to enterprise AI solutions and complex digital products, including agentic systems, platform architecture, super apps, loyalty programs, customer data, and governance. Organizations bring me in when projects are technically complex, delivery is stalled, or a fixed deadline requires decisive action.
If you are managing similar programs, I welcome the opportunity to connect.
Education, certifications and languages
- Advanced Technical Diploma in Computer Application Development (Spain)
- Business Strategy and Digital Innovation (ThePowerMBA)
- AI Product Strategy for Leaders, and AI Product Leadership (Product Faculty)
- Design Thinking (IDEO)
- Agile Scrum, SAFe, Design Sprint
- Languages: Native proficiency in Arabic, English, French, and Spanish
How I Think, Build, and Lead
My favorite quote:
“Everything around you that you call life was made up by people that were no smarter than you.”
I like this quote from Steve Jobs because it instills a sense of agency. Agency begins when you recognize your ability to take control and make independent progress toward your goals.
Observing any man-made environment, such as roads, signs, apps, payment systems, buildings, or institutions, reveals that everything was designed and built by people no smarter than you. Recognizing this makes the world feel less fixed. With focus and time, you can learn how systems work, why they fail, and how to improve them.
This perspective is especially important in the AI era, as work, products, and industries are rapidly evolving. While we cannot control when disruption occurs, we can choose to understand, build, and adapt rather than remain passive observers.
My favorite frameworks:
- First principles thinking: Used to break problems down to their fundamental facts and distinguish what is true from what is inherited.
- Wardley Mapping: Applied to identify sources of advantage in a value chain, track their movement, and inform build-versus-buy decisions.
- AI evaluation frameworks: Used to assess RAG quality, model selection, guardrails, orchestration, memory, context, human review, observability, latency, and cost-to-serve tradeoffs.
- Hamilton Helmer's 7 Powers: Used to separate a good feature from a defensible business, especially in AI markets where model-level advantage commoditizes quickly.
- Netflix's Context, Not Control: Used to scale decision-making without turning leadership into approval management, giving teams enough context to make high-quality decisions autonomously.
- Anthropic's AI-native product development: Used to build close to the model, prototyping before over-planning, and letting real model behavior shape workflow design.
Books that shaped my thinking:
- 7 Powers: The Foundations of Business Strategy, Hamilton Helmer
- No Rules Rules, Reed Hastings and Erin Meyer
- Essentialism, Greg McKeown
- Playing to Win, A.G. Lafley and Roger Martin
- The Phoenix Project, Gene Kim, Kevin Behr, and George Spafford
- AI 2041, Kai-Fu Lee and Chen Qiufan