Unleash the power of AI

into your business

Our algorithms are designed to learn how to solve a problem by themselves. To perform this task, they look at examples and adapt themselves accordingly, a process commonly called machine learning.
Our technology is primarily based on deep neural networks, a rapidly-evolving research field which draws inspiration from the way the human brain works. Thanks to the hard work of a vibrant community of software developers and data scientists, deep neural networks have managed to achieve state-of-the-art performance in a variety of human-level tasks.
Put simply, our algorithms are a powerful tool of statistical analysis. After looking at your data, and regardless of its complexity, they will discover subtle relationships that would be challenging for a human to discover.
However, what deep neural networks are not is a one-fit-for-all magic answer to any possible question. Every problem requires a tailored-made approach. Our job is therefore to design the unique solution that will meet your specific requirements.


Sebastien is the lead data scientist of Vejja. A graduate from Ecole Centrale Paris and Sciences Po Paris, he first worked in finance before dedicating himself on what he likes to do: solving mathematical problems with computers. If you want to talk about Restricted Boltzmann Machines, Gradient Descent Optimization or Random Forests, he will take your call free of charge.


Mehdi is our big data architect. Mehdi likes nothing more than parallelizing matrix-vector operations in high dimensional space across data clusters. He is also the founder of Smartpoint, an IT venture specialized in business intelligence services. Mehdi graduated from Ecole Centrale Paris.


Guillaume is the business developer of Vejja. He is a highly-experienced professional coming from the insurance industry. As all serious law school graduates, Guillaume can only drink vintage Bordeaux red wine. Anything else will turn him sick.


Thomas is one of the elite developers admitted into Ecole 42. We call him "the genius". If you need to communicate with him, please write to him in Python/C/ASM/any other computer language of your choice. Please do not forget the semicolons at the end of each line, otherwise he can only reply 42.

Our journey in artificial intelligence first started when we were developing software for the automated classification of documents. In order to tackle this challenge we had to resort to find innovative technologies and started implementing unsupervised Boltzmann Machines. We were quickly surprised by the power and capabilities of deep neural networks and decided to test their potential in other domains, including image, speech and text.
Throughout this journey we've developed a number of commercial applications for this technology. Below are examples of what our algorithms can achieve.

  • Understanding client requests in natural language. Customer service desks around the globe routinely process million of client requests each day. Wouldn't it be great if we could automate this task with a little help from artificial intelligence? To reach this goal, we use natual language processing, a technology designed to understand langage as it is spoken and written by humans. The algorithms examine the structure of input text and characterize its contents.
  • Targeted promotional campaigns. Promotional campaigns can be a waste of money if they are not targeted at the right audience. We use a sophisticated combination of predictive tools to crawl through the purchase history of client databases. With enough datapoints, the algorithm is able to spot out which client is likely to be positively impacted by a specific marketing operation. And one day, the financial director asked how much revenue each of your marketing dollars would generate - and you knew the answer.
  • Automated chatbot. One of our most ambitious and innovative projects to date is to build an automated chatbot that can mimic the interaction between human beings. We use state-of-the-art research in recurrent networks and reinforcement learning to reach this goal. This project is currently under development.

If you want more information about our activities, please use the form below to contact us.

Click here to send us an e-mail