I'm Edwin Helet. I built payment systems for European banks before training in strategy at Emlyon and CentraleSupélec. Now, I would like to advise companies on their AI and data transformations, from strategic framing through to what actually runs in production.
Strategy wasn't what drew me in at first. The engineering was, a childhood ambition to build useful things. Working on real systems, under regulatory and delivery pressure, is what shifted my ambition: from "designing" to "understanding the needs technology has to answer". The specialised master's formalised that shift.
I worked inside a genuinely constrained environment: SEPA and card payments, thousands of daily transactions, European regulation setting the deadlines. You don't improvise there. That experience is what lets you know, before a strategic decision is made, what will actually hold downstream.
A hybrid profile isn't about ticking two boxes. It's about drawing a direct line between a business stake and the technological lever that answers it, and above all about choosing between levers when every one of them looks defensible on paper.
A dozen technical and business projects across engineering school, the specialised master's and entrepreneurship. Alacasa and CityZN took me across the entire value chain, from market positioning down to the database.
A transformation is never decided on a single segment. Strategy without delivery stays an intention; delivery without strategy produces tools nobody uses. Every skill below carries its origin, because expertise without provenance is just a claim.
Designing, building and hardening what has to run in production. This is the ground where a recommendation is proven, or falls apart.
Distributed and cloud architectures
Automation and software quality
GDPR compliance and open banking
On-device AI and local processing
Artificial intelligence, computer vision, statistics
Reading a market, choosing between options that all look defensible on paper, and turning a finding into a business model that holds.
Market research, segmentation, competitive analysis
Business model and value proposition
Business cases and AI-augmented models
Management control and performance management
Competitive intelligence and geoeconomics
Framing it, bringing teams along, and seeing it through. This is where most transformations are lost, and rarely for technical reasons.
Agile at scale: SAFe, Scrum, V-model
KPI-driven steering
Field validation and stakeholder interviews
Framing AI use cases
Product roadmap and technical trade-offs
Emlyon Business School × CentraleSupélec
Strategy, management control, performance management, B2B marketing, negotiation, competitive intelligence, geoeconomics, international law. Academic exchange at CFVG, Hanoi and Ho Chi Minh City.
Alacasa SAS
Technology, product and AI for a platform that inverts the real-estate brokerage model: index qualified demand rather than listings.
SBS (formerly Sopra Banking Software)
Digital transformation for two European banks (ASN Bank and Argenta) across SEPA payments and Visa/Mastercard cards. 40 sprints, 6 SAFe program increments, an ecosystem of over 50 microservices.
Université de Sherbrooke
Five months of experimental research: trials across dozens of materials, validation prototypes, data analysis and technical recommendations.
EPF Engineering School
Generalist training then specialisation: artificial intelligence, cybersecurity, computer vision, DevOps, cryptography, statistics, design thinking.
Each case study follows the structure I apply on engagements: context, diagnosis, levers, trajectory, results. The format is part of the argument.
Two European banks, more than 50 microservices, and deadlines set by the regulator, not by the delivery plan.
200+
Index qualified demand rather than listings, in a market where 80% of the leads an agency receives are unusable.
80%
A three-month innovation challenge run by a team of eight. Three prizes out of four, and a value proposition the field rewrote.
3/4
A personal finance app built so the data goes nowhere. Local processing, no account, no tracker.
0
Consulting project for Onepoint (2023–2024): analysis of an infrastructure availability measurement problem using Raspberry Pi micro-sensors.
AI isn't primarily a business topic: it's the accelerator of issues that reach far beyond it (governance, sovereignty, the environment) that organisations inherit whether they saw them coming or not. Four convictions shape how I approach them on engagements.
Substitution, role transformation or augmentation: the effect is never the same depending on the task, and conflating the three means getting the support plan wrong. The real barrier to adoption is rarely technical. It's trust, and trust is built through explainability: being able to say why the system decided what it decided.
Knowing who decides, who validates and who is accountable has been a corporate governance question long before AI: finance, compliance and data security already demanded it. What changes with AI is the speed at which a poorly governed decision can propagate. The AI Act formalises this requirement for high-risk systems, but the real issue stays organisational before it's regulatory.
Depending on a foreign provider to host your data didn't start with AI: GDPR and the debates around the US Cloud Act had already raised it. Generative AI has mostly concentrated the dependency further, with a handful of players now controlling both the compute and the models. That's what's fuelling the emergence of a sovereign market in France and Europe (OVHcloud, Scaleway, Mistral), less out of ideology than a need to keep an exit option.
Digital's energy consumption and non-financial reporting obligations existed before generative AI entered daily use. Training and inference at scale have mostly made that cost more visible, to the point of sometimes contradicting the very sustainability commitments a company has made elsewhere. Tomorrow's real limit, then, won't be AI's ability to replace people, but the cost, energy as much as financial, past which it stops being worth doing.
I am based between Paris and Lyon, available for consulting engagements in AI strategy and organisational transformation. The simplest route is to email me directly.
[email protected]