Sites Inria

Version française

Start-up - Sailendra

Interview with Régis Lhoste

Régis Lhoste Chapo

At a time when Web content suppliers are facing profitability problems, Sailendra has developed a software platform making it possible, through the analysis of uses and the anticipation of user actions, to target a customer base and offer information likely to interest it. The recommendation engine marketed by the start-up has already been the subject of several contracts.

What is the history of Sailendra's creation?

Régis Lhoste:  Anne Boyer1 has been working within the MAIA team since 2001 on an information recommendation system for news systems, which were in their initial stages. She continued the research work with Sylvain Castagnos², a young PhD student at the time, when I left Inria to launch an engineering career. When they were creating the start-up, they were looking for a person with experience in the industry and contacted me in 2008 to offer me the position of director. I took control of the company rather naturally, as I was familiar with the technology. This form of technology transfer was more interesting for us than licensing.

Where does Sailendra stand currently?

Régis Lhoste:  Today, Anne Boyer and Sylvain Castagnos are still researchers within the KIWI team of LORIA3 but are also shareholders of the company and part of the scientific board. They are responsible for R&D in terms of algorithms to contribute new advances to the technology developed by Sailendra. The company, which has three full-time employees, is housed in the technology transfer branch of Inria Nancy-Grand Est on the LORIA site. Inria was essential in the creation process by giving us the means to start up a high-tech company under ideal conditions in terms of human and physical resources. Regardless of the future form of the company, we cannot do without the scientific advice provided by the researchers.

Precisely what is Inria's role in the start-up?

Régis Lhoste:  The recommendation of information based on uses is a rather new field and therefore has few players. With the big concern related to the Internet today being the judging of the quality of recommendations, being in contact with researchers allows us to have the top-performing, most advanced technologies and therefore have the top-performing recommendations. We offer a solution on a case-by-case basis, which is possible because the researchers help us to analyze the customer's system and adapt us to the specific features of each one. The Inria calling card is important. Of course, the institute does not search for markets for us, but its network makes it possible to direct clients looking for recommendation solutions towards Sailendra.


1- Anne Boyer: computer science professor at Université Nancy 2, manager of the Kiwi research team. Responsible for the design of future generations of products.

2- Sylvain Castagnos: graduate of ESIAL, PhD in computer science from Université Nancy 2. Responsible for the modeling of community behaviors.

3- LORIA, (Lorrain laboratory for computer science research and its applications) is a Joint Research Unit for several institutions: Inria, CNRS (French National Center for Scientific Research), INPL (Institut National Polytechnique de Lorraine), UHP (Université Henri Poincaré, Nancy 1), and Université de Nancy 2.


The Sailendra  software platform offers a collaborative recommendation engine, named as an analogy with search engines. A result of research conducted within LORIA, this tool is made up of software modules that associate social navigation  (exploitation of information about a user to deduce his or her preferences, habits, and tastes) with collaborative filtering  (electronic word-of-mouth on the modeling of a group's behavior and the observation of its activities in order to deduce its future actions).

The Sailendra software is intended for online merchants or operators of intranet or document sites wishing to personalize access to their information sources and enhance their services (creation of virtual communities, relationship building, targeted marketing, etc.). Configurable depending on the company's context, the platform complies with the CNIL recommendations regarding privacy in the sense that no identifying information is collected about the user without his or her knowledge.

The start-up, which has identified two potential competitors in Europe for now, can deploy these services for hundreds of thousands of users in real-time in the field of e-commerce, online news sites, mobile telephone or interactive television services, searches for information on a document intranet or in a digital library, etc. The sale of Sailendra modules is done in the form of license revenue based either on the revenue generated by the recommendations (commercial sites) or the number of end users of the product in one year (intranet or document sites).

Keywords: Sailendra Régis Lhoste Start-ups Search engine Platform