As a scientist, entrepreneur, and software engineer I help European businesses and institutions with risk management intelligence.
Often, risk has monetary consequences, e.g., payment fraud or failed package delivery; however, risk can sometimes present gain opportunities or health hazards including death. I work on three problems: payment fraud, triage optimization, and time to failure.
Processing payments online goes right most of the time, but it fails in some scenarios, e.g., if a buyer abandons his cart, a technical failure occurs, or a payment is fraudulent.
In cases of payment fraud, the total costs are associated with chargebacks, lost sales, staff time, and technology needs. These costs represent 5 to 15% of an e-commerce service's net margin. To prevent payment fraud, e-commerce services use systems with lists, rules, risk scores, and self-learning algorithms, such as Oriskami.
After payment, e-commerce services pass prepared orders to shipping carriers. Total shipping costs include delivery costs and costs associated with fulfillment issues or goods damaged or lost during delivery. These costs represent 5 to 10% of an e-commerce service's net margin.
Shipping carriers compete based on the geographical area they serve but also on the size, weight, and type of packages they handle. To reduce the total shipping costs, e-commerce services can match orders with the most competitive shipping carrier in terms of price, delivery time, risk, or a combination of these factors.
Time to failure
10% of 60+
In public health, heart failure is a chronic disease that requires lifelong treatment and affects 10% of the population aged 60 or older. Some treatment strategies are coronary bypass surgery, ventricular assist devices, and heart valve bio-prostheses.
To study the performance of these bio-prostheses, heart surgeons monitor reintervention rates due to heart valve dysfunctions, which are analyzed as the time to failure in statistics.
Know-how and approach
Using scientific and technical expertise as well as key principles, I unlock how risk is managed in businesses and institutions over the short and long term.
Statistical learning helps us discover similarities in data and categorize data. I use it to break down problems, model text data, or reveal structure.
Epidemiology helps us design studies and collect, analyze and interpret data. I use it to design experiments, impute missing values, compare groups, or measure time's influence.
Simulation helps us imitate real-world processes with algorithms or statistical models. I use it to mimic live systems and introduce and back-test changes.
Financial risk management helps us identify sources of risk, measure risk, and develop plans to address risk. I use it to value the risk cost, estimate risk exposure, or set priorities.
Visualization let us use graphics to help people understand, build hypotheses for, and rationalize problems. I use it to explore data, reveal multi-dimensional relationships, or summarize data.
Statistical programming helps us develop, test, and operate systems involving statistics. I use it to explore, prototype, and make statistical models.
Full-stack web development lets us build web apps served by remote systems. I use it to make accessible, install-free, update-free, centrally managed software.
Database design helps us store, retrieve, organize, and inter-relate data. I use it to setup business processes or make in-database calculations, thereby avoiding the cost of moving data in and out.
Virtualization helps us abstract operating systems and networks. It provides application isolation, better resource usage, and cost savings. I use it to build small versions of enterprise systems.
Effective communication lets us send a message, receive it, and act upon it. I believe candor, honesty, and integrity help us best achieve effective communication.
I feel that care and empowerment for the people I relate to helps build engagement and reciprocity. I also understand that sharing a vision helps everyone align towards one goal.
To reveal and go past existing frontiers, I usually source process inputs to understand their bits and pieces, and, quite naturally, grow in ownership and work autonomously.
The following are some of my professional contributions:
Moving the needle in business
For an online travel agent with annual sales of two billion euro, I helped save 60 million euro in sales over four months by enhancing how the company managed payment fraud.
Working with an e-commerce service's support team, I setup risk management processes that reduced chargebacks by 10x over two nine-month periods.
Over four consulting days, I proposed changes to help a business cut its shipping costs by 30%.
Over the past 15 years, I created a French and a Dutch consulting companies, a Delaware-based C corp, a French SAS company, and an Estonian limited company.
I attended EuraTechnologies' incubator and accelerator in Lille, France, where I was awarded an entrepreneurship grant there as well as France's "young-and-innovative company" status.
Background in a snapshot
I received an electrical engineering degree and an M.Sc. with a focus on computer science, in France, received a Ph.D. in statistical learning from the University of Leiden, in the Netherlands.
Then I worked as a postdoctoral scholar in human genetics at the University of California Los Angeles, in the USA, as well as in statistical epidemiology at the Leiden University Medical Center, in the Netherlands.