While waiting for the announcement of the names of the new startups making up the next season of IONIS 361, the incubator for students and graduates of the IONIS Group, ESME Sudria invites you to rediscover one of the companies from its very first season: Culmineo. Founded and run by a graduate of our school by the name of Christophe Desnoyer (ESME Sudria Class of 1993), this IT startup provides its clients with innovative technology to improve the performance and availability of their websites. We took the chance to revisit the career path of its creator, starting from the time he graduated from ESME Sudria.
What have you been up to since obtaining your engineering degree?
When I was a student, the school only had two majors: electrical-mechanical and electronic engineering. I went for the latter and therefore became an “electronic-mechanical” engineer with a specialization in computer science – for which, incidentally, I graduated at the top of my class. When it came time to look for my end-of-studies internship, my average grade was exceptionally high – I only ever received one, an A+! –, which allowed me to be taken on by a company known to be extremely selective in its choice of interns and employees: Dassault Electronique. My internship went well and I was subsequently hired as a full-time employee. Thus I began my career in the defense sector. I stayed at Dassault Electronique for six years, gradually climbing the corporate ladder from Software Engineer to Manager of the Software Engineering Tools and Workshops Department. In 2000, I decided to shift both “course” and sector by becoming Technical Director at Galileo, a company that supplies reservation systems for the tourism industry. In 2005, still working in tourism, I was hired by Groupe Nouvelles Frontières/TUI as Group CIO. A year later, I continued along my path in this sector by joining the IT Management team of the Accor Group, first as Reservation Systems Director and then as CIO and Financial and Administrative Director. And then, in 2012, I was struck with the urge to change my life completely and start my own business.
You have experience working for large companies, both in the defense and tourism industries, and have held high-level positions. You even seemed comfortable with this “routine.” Why the decision to launch out on your own?
It was precisely this routine that made me take the leap. I felt as if I had been flirting with the idea for a while, especially considering that in IT, the only position higher than CIO is CEO. However, my goal was not necessarily to become the CEO of a large company. As CEO, you deal with a lot of politics and inactivity, and little real innovation. Personally, I was too attached to the “hacking” side: I was a CIO who understood technology – which today remains quite rare – and I have always loved to write code. I wanted to rediscover this taste for invention and innovation. Furthermore, when you reach the age of 40 and realize there’s something you haven’t tried yet, you tell yourself it’s now or never. And since that something meant entering the domain of digital startups, where innovative ideas take center stage, I knew it was the ideal moment because I just happened to have a good idea. I first came up with this idea while working as Reservation Systems Director at Accor. I identified a specific problem, then racked my brain until I found its solution: that is how the Culmineo concept was born.
That proves there is no age for starting one’s own business.
True – and I know other people who got started even later than I did. On the other hand, it is also true that starting one’s own business is extremely time-consuming and stressful. It requires an investment beyond that which is requested of a “simple” employee. You have to know exactly what you are getting into before taking the plunge.
How did Culmineo come into being?
Although the company was created in 2013, it required a rather long incubation period as the solution had to be developed in order for the product to be completed. That takes time, especially if you are working alone. I spent a lot of time thinking to come up with just the right solution to the problem I had encountered at Accor. Once I finally found it, I implemented it, patented it, and was off and running.
What was the problem?
The problem deals with risks to traffic on the Web. When you have a website, you also have infrastructures and a limited number of processing capabilities. But most of all, you have the entire world at your doorstep. If suddenly the whole world feels like connecting to your site at the same time, then your site is going to crash. We can all think of examples of sites that have crashed when they were launched, while they were trending, or during a seasonal event such as the beginning of a sales promotion on an e-commerce site. This is a problem that even affects major websites in possession of an exorbitant amount of resources. It also concerns peaks in traffic which are not connected to normal activity but to cyber-attacks such as DDoS attacks. I wondered, therefore, how a site can survive when it is faced with a massive wave of traffic, all the while taking into account the fact that such a wave can include cyber-attacks as well as real customers or users. It is a genuine problem because DDoS protection mechanisms, for example, block traffic indistinctly, without considering these real customers who wish to remain on a site.
The gist is to separate the wheat from the chaff…
Exactly, all the while regulating the wheat! Think of a website like an interstate entrance ramp at rush hour: all the cars are “wheat” and wish to enter the highway at the same time. Except if we let them do it, the result will be an instant traffic jam. To avoid such jams, it is therefore necessary to regulate in advance, funneling them through bit by bit and as quickly as possible. Although this is a problem any IT infrastructure manager may encounter, the market currently lacks reliable solutions for treating it. Existing technologies are unable to differentiate fake queries from real ones. When activated, they become veritable open gateways for DDoS attacks. This means a savvy attacker can pick up on the regulation mechanisms which are employed and subsequently bring down the targeted website more easily. Incidentally, one of the mechanisms used by Accor during my time there worked in this manner. That is what inspired me to find this new solution.
When did Culmineo begin commercializing its technology?
First of all, we had to design it, simulate it in the lab, and verify its control algorithm. We then needed to secure the concept by filing its patents, and implement the technology by developing software that could withstand such a considerable load. Finally, we had to find customers with a traffic problem who wished to use our proof of concept to resolve it. For this final step, I was able to count on two internationally-renowned groups who as it happens had a similar problem when trying to handle peaks in traffic and improve the performance of their websites – regulating traffic also involves maintaining performance at an optimum level. To sum up, the development of the technology lasted more than one year while the validation period of the proof of concept in the field with real customers took place between the end of 2014 and the end of 2015.
Why did you decide to join IONIS 361?
To prepare for and proceed to the “next level,” namely that of commercialization. By integrating an incubator like IONIS 361, we can participate in an environment that allows for contact with business and engineering school students who may one day wish to join our team. It also provides us the chance to communicate with other entrepreneurs and organizations to establish contacts and eventually form partnerships and collaborations. Because Culmineo mostly targets B2B and (especially) large groups, the mentors of the incubator are potentially extremely important contacts.
Check out Culmineo on its website and LinkedIn
Big Data holds no secrets for Erwan Le Covec (ESME Sudria Class of 2016). One of the school’s recent Symposium Prize winners for a project based on data processing, this Big Data and Digital Intelligence student remained in his chosen field by taking his end-of-studies internship at Hewlett Packard Enterprise (HPE). A few days after completing his internship, he reflects on his fascination for this rapidly growing field.
Why did you want to attend ESME Sudria?
At first, I chose the school for its Energy major. I was extremely interested in the major, and knew ESME Sudria had an outstanding reputation in this domain. As I made my way through the curriculum, however, I realized that the major was not for me. Instead, I decided to try out another area that I enjoyed, namely information technology. I therefore switched to the Big Data and Digital Intelligence major.
What makes IT so interesting?
I’ve always liked it, maybe because it’s a field where you really have to dig in and “get your hands dirty.” With code, you can create something interesting from nothing. That definitely has its allure. Finally, IT is the gateway to new technologies: there are always new things to see, learn, and discover.
You took an internship at Hewlett Packard Enterprise (HPE). How did you find it?
It all started with my 5th year project, which I completed in partnership with HPE. It was about WADA, a project that analyzed weather and health insurance data. Because this project went so well, the company asked my colleague Anthony Ginoux and me if we’d like to do our internships with them, in the same field and still using HPE tools, but on another subject.
What was the subject?
It involved a study of the world’s refugees. In the earliest stages of the project, our mission was to create a demo with HPE’s Big Data tools. As we were pretty much free to choose the subject for this demo, Anthony and I gathered quite a bit of open data to see what might make for an interesting topic. That’s how we stumbled upon the data on refugee streams. Once we realized a lot of relevant data could be tied together, we were off and running.
What kind of data did you gather?
Well, there was the data on the refugee streams, as well as data on wars and on natural and industrial disasters throughout the world. We also gathered data on religions and development indicators such as a country’s GDP and birth and mortality rates.
How does one get hold of open data?
Simply by combing through open data sites like those of universities or international organizations. Incidentally, we found the refugee stream data on the United Nations website, which has quite a few records.
Big Data not only makes it possible to analyze correlations between more factors, but also to use logic to make predictions. Were you able to make any predictions?
That’s exactly what we tried to accomplish: predict refugee streams in the future. By using a variety of algorithms, we succeeded in coming up with something rather compelling. In our final application, called Refugees Flow, you can choose a country on a map and see the number of wars and disasters that are estimated to occur between 2015 and 2030. Consequently, an estimation is provided of the number of refugees who will flee the country during any given period.
Does the application cover the entire world or just a few specific geographic areas?
Since data was lacking for certain countries, our predictions are only based on half of the world’s countries.
Will this application be made available to the public?
In theory, only HPE can use it. That said, the company will release some non-modifiable graphical visualizations of the results.
What are your plans now that you have finished your internship?
HPE offered to keep me on, but I decided to look elsewhere as their offer didn’t quite match what I was looking for. During the internship, my work was extremely close to that of a data scientist: this is the direction I saw myself taking. I therefore simultaneously began to look for opportunities to work abroad as part of a VIE (International Volunteer Mission in a Company). I found one in Brussels and applied: it’s a one-year VIE (with the possibility of renewal) as a Junior Data Scientist at Keyrus Biopharma, a consulting firm in the pharmaceutical industry. Keyrus Biopharma has accepted my application, so now I am just waiting for approval from Business France, the organization in charge of VIEs, before going.
The data scientist profession is what you might call a “new profession.” What do you like about it?
The domain of Big Data is constantly evolving: there are always new products, new technologies, etc. It requires the continuous search for and testing of new methodologies. I am especially drawn to the “exploratory” side of the profession: data needs to be explored continually if you wish to reach worthwhile conclusions. You feel just like an explorer of modern times.
Why did you decide to go abroad?
Because I also love exploring the world! I get a thrill from discovering new cultures so really felt the urge to go abroad. Admittedly, Brussels is not much different from France. But one day I hope to work in Singapore or Hong Kong, two cities which are known to have a large number of data scientists. After a year or two in Brussels, I’ll have the necessary experience to apply for jobs in these cities.