Raphaël Géronimi

Raphaël Géronimi

Founder, ArsChitectura

After training in mathematics at École Polytechnique, I spent 14 years in Silicon Valley, at McKinsey, and at Société Générale before founding ArsChitectura to bring rigorous, explainable Machine Learning and AI to high-stakes industries over the following 14 years.

Why ArsChitectura

After co-founding a startup in Silicon Valley, advising executives at McKinsey, and managing the data program at Société Générale, I found myself spending far more time in organizational meetings than doing the work that had drawn me to this field in the first place: mathematics. Meanwhile, colleagues in e-commerce and ad tech were experimenting with the latest machine-learning techniques coming out of Google and the real-time web - innovations that large corporate environments still have not adopted.

In 2012, I stepped away to refocus on this opportunity. The idea behind ArsChitectura was simple: combine deep mathematical expertise with the business and program-management experience I had accumulated, and offer that combination directly to companies tackling hard problems on large-scale data. The first client fell through before I started. Two new engagements arrived almost immediately, and the practice has grown from there.

What I have built over the years is a bridge: the ability to sit with a leadership team, understand the business stakes, and then go design and deliver the mathematical solution myself or with their team - whether that calls for a fully transparent model that regulators can audit, or a high-performance system serving hundreds of thousands of predictions per second. There is no single technique that fits every problem, and knowing which approach a given situation demands is itself a skill I have built over years of working across very different industries and contexts.

What I Do

I have implemented a broad array of innovative projects including:
  • Fintechs - Coached and led mathematical breakthroughs for
    • A $3 billion consumer lender (credit scoring, fraud detection, and AI automation)
    • The world's leading software provider for financial institutions (error detection)
    • A leading automobile insurer (risk pricing)
  • Web advertisers - Delivered machine-learning solutions for
    • Real-Time Bidding (>300,000 online predictions per second)
    • Search-Engine Marketing (prediction models and portfolio optimization on >1 billion items, AI optimization)
    • Search-Engine Optimization (semantic matching models, repulsion models)
    • Recommendation (graph models, ranking models, uplift emailing models)
  • Venture Capitalists, Founders, and Academics - Advisor to
    • A US & Israel VC fund (15 deals advised)
    • A few startup companies

My Background

Education

More...
◄ Return to homepage