Good morning, my name is Stéphane Mallat, I am a former student at the Polytechnic School. Afterwards, I went to the United States to do a thesis, there, at first, I thought of coming back to France soon to join a company, and I was bitten by the bug, I had to research whilst working on this thesis, I became therefore Professor at the New York Courant Institute, in the university of New York. And there, whilst researching, something struck me, most of my colleagues, in those days and it was always the case, as well as their research activity, very often, created companies on the side, even became entrepreneurs or went back to the academic field. Therefore, in those days, creating a company was for me something totally natural, even though we were in academic surroundings. I then worked in a field which literally skyrocketed, which is a field of applied maths called the wavelet transform, and in those days, in the 90s, on of the things which really hit me, was that although much of this research had been made in France with the help of totally remarkable mathematicians, the majority of the companies, even all these startups were created in the United States. Therefore, when I came back to France, in the year of 1995, I came back as a Professor at the Polytechnic School, one of the ideas I had in mind, was that if a new opportunity arises, then at that time, why not try in France? And that is what happened. It was at the start of the years 2000, at the time of a thesis which has been done by one of my former pupils, Erwan Le Pennec, who did an extremely good job on the building of orthogonal basis. Then, the idea was that, we can look at images as points tables with structures. One of the very important information in these images, are the geometrical structures. The images can be seen through the scales. That is the idea which brings this wavelet transform. And at the same time, we can also look at them from a geometrical point of view, and the question is: can we combine the two points of view? That is the question that Erwan Le Pennec tackled in this thesis, through the creation of these orthogonal basis of bandlets. So what are those bandlets? They are both multi scale structures, which can then be very small, even big, but which at the same time marry the edges of the images and then allow to represent these images in very parsimonious way through the linear combinations of these few bandlets. So, from there, and basic results obtained by Erwan Le Pennec during his thesis, we told ourselves that there would be potential applications in image processing, especially for making cuts or to improve the image quality. And there again, this question arose. Why not do it in France after all? So far everything was done in the United States, why not launching a startup in France? And there there was very favorable set of circumstances because I had two former doctorate students, Christophe Bernard and Jérôme Kalifa, they, for different reasons, both felt like creating a startup. In those days, Jérôme Kalifa was post-doctoral in the United States, Christophe Bernard was teacher-researcher at the Mines school, so the four found each other and we said to each other, why not take this opportunity to create a company? So, behind the creation of this company, there was a dream, it was to create the new Microsoft image processing. It was really about creating something that would completely skyrocket. We did not at all feel like creating a small utility company, which would, somehow, would prolong our activity in the laboratory because the research, it can be done in a laboratory, the idea was really to create something would could skyrocket on the international level. And behind all that, there is this small researcher dream which is to have a deep impact using tools we achieve. As a mathematician, we develop theorems, eventually small applications, but that remains quite often restricted to some colleagues around us or in an international community. There, the idea, one of the dreams behind all of this, was to follow a theorem and to potentially get a product that was ignored at the time, and which could eventually have an important impact. Therefore there was a small egocentric side behind all this : could we have an impact on the world through a theorem? So, the first step, obviously, was to create this company. The first one who launched himself, full-time, was Jérôme Kalifa. His salary obviously had to be financed. And the funding for the startup was first of all made through a service activity. The first contracts we had were with the CNES, with large companies like Alcatel Espace, with whom we had already relationship in technology. At the same time, we had a large help from public powers, that is, at the time, was called Anvar and then became Oséo. And it was done in two steps. First of all, at the time of the incubation of the company, and then we received the first national prize for the creation of the company, which has been very important for two reasons, obviously, from a financial point of view, with help, but also because, all of the sudden, there is an outsider telling us : what you are doing, it is not just anything, actually, it is possible to redo it. Therefore, that presented a sort of reinforcing which has been very positive psychologically. Additionally, at the time of preparation of the file. It was done with a consultant, of great talent, by the way, called Karine Cheltiel. It allowed us to think about the company for the first time. In other words after six months, what will potentially become of us in one year, three years, five years. And this was a first contemplation process a little more in depth about our future which was very important, that we were able to develop throughout this competition. So this was the very first phase and then, the company slowly grew through a service activity. This service activity played two roles, a first financial, of course, in other words allowing us to develop the technology, but also a learning role. Throughout this activity, we came out of the laboratory and maybe the first thing we started to learn was the job of engineer. There is a big difference between the engineering and the research jobs, an engineer must work in a very limited time, in three months, and in three months eventually resolve a very complex problem, whereas the researcher is potentially going to tackle the same type of subjects, but over three, five, ten years. So the procedures are not at all the same, and we learnt that, by working in a limited time. The second thing we have started to learn, but only at the beginning, we started to listen to the clients. We needed a tremendous amount of time. As a researcher, we are often used to impose or to want to impose a vision to the rest of the world. It is the reason we can eventually develop approaches which are entirely original. Then it was necessary to listen. At the beginning, our vision was: we have a technology which allows to represent in a more parsimonious way the images, well we are going to compress and we are going to do it on the Internet, on cameras, videos and so on. And quite a lot of time was needed to understand what the clients were telling us, that is the compression, it is very good, maybe you would perform better, but it does not bother us because there are standards and it is not where the value is for us. Where there is value, according to us, is the ability to improve the quality of the images, whether it is true for the satellite images, we worked therefore with Alcatel Espace or with the CNES. For geophysics, we worked with the Americans, BHP Billiton, or infra-red images, and so on. So practically all the requests were about the image quality whereas we used to arrive telling people : well here we are, we can compress, what are you offering? Therefore, the development of the company was absolutely not linear. To tell the truth, we have drawn four completely different business plans. We started, therefore, with a service activity which, from the beginning, was clearly for us just a startup phase knowing that what we wanted was to develop, at all costs, a product and if possible internationally, on the biggest possible market, as I mentioned it at the beginning. So, at first, we thought about a compression product and then we realised that the generic compression was not, possible, so we wanted to develop compression for passport photographs, identity card, and so on. That was the training of the strategic marketing. We finally realised that this market was far too small and, eventually, destined to die out because of the memory price declines. And we did a 90 degrees turn, we turned around, we realised that, it was 2004, that it was the arrival of the high definition flat screens, while the video content was in low definition, and then the problems, at the time, was to capture these low definition and resolution images, and artificially, by mathematical calculations, make them high definition and resolution images. And therefore, we decided to launch ourselves in it and then, obviously, straight away, it was a big jump as it was the way from soft to hardware. All this must be made for real on videos, and it was a jump as it was no longer compression but restauration- improving as much as we can the quality of the image. The extraordinary aspect of maths, is once we have generic tools, they can be transposed. We then knew the types of technic to be able to do this in a complete naive way, because at the time, we thought that the hardware, the software, all, was quite similar. Therefore, we decideed to launch ourselves with first demonstration videos to raise funds. We raised funds, and I will come back to this And the first step was dedicated to what we call the broadcast market, in other words the professional market, developped the video conversation box for television channels, for the positions of producers, and so on, so professional market. Which allowed to obviously sell, the equipment at higher prices with the ultimate aim to enter the high definition television market, in other words sell microchips for flat screens. And that is a market, was a market, and still is, a very large market in the region of 2 billion euros. And this market is, of course, a market, with major stakeholders such as Samsung, Broadcom, Intel, and so on. So it was our objective. In the meantime, we realised that we had to choose. A startup has enormous difficulties following two paths at the same time, because it is necessary to be extraordinarly fast with important competitors. So, the last step of the company, is to totally focus on this high definition television market, in other words design, build microchips and commercialise electronic microchips for the high definition television. At the time, it was in 2007, we found ourselves in a situation when we had to choose. Either do a second fund raising, knowing that a semi-conductor company like this one, typically, must raise in the region of 50 millions euros before reaching a balance. Therefore, the next phase, was to raise funds in the region of 15 millions euros, or to be sold. At the time, it was clear that a phenomenon of reaggregation of the consolidation of industry existed. The margins were declining considerably. The only ones which are currently continuing to thrive and this even more the case now, are the companies in Asia, in China, Taïwan, to the slightest extent in Japan, Korea of course. So the question was to know in five years time, would the company, Let It Wave, be one of the biggest largest in the world or not. The answer was quite obvious. It is No, we were far too late. And then, this phenomenon of concentration, evidently, was going to be the end of us. So we decided to try and sell, knowing that at the time, we had almost the best technology on the market. It was known by our clients, it was known by our competitors. So it was the best time to sell before all the development phase of commercial support which is extremely heavy on the industry of the semi-conductor for televisions. So we bypass this selling phase, in other words, we sold the company in June 2008, and it is how the story ended. Therefore, when we decided on 90 degrees turning point towards high-definition television and hardware, it was obvious that very large capitals were needed, as to build only even one ASIC it is necessary to have in the region of 2 millions euros to do what we call masses, the negatives. So, it is an industry which requires a lot of capitals, and for the first phase we needed about 6 million euros. Therefore we raised funds in a classic manner, we went to see all the investors in the square, over France, and we found two investors we left with, and who were Iris Capital, the main investor, and E Source. Therefore the relationship with the investors was a success. At the same time, I think that it is important to understand well what a relationship with an investor entails. A relationship with an investor, firstly an investor, what we are looking for, is money. And eventually, they will tell you that they will not be able to give you advice and so on. But above all, what interests you, is money. It brings you some kind of rigor, as it is necessary to justify the path we are on and the future perspectives. Then, it requires indeed working on our conscience and to ask ourselves the right questions, but they will not be pas a priori in general the people, obviously, who will run the company. They are the people who are going to be on the outside and who are going to check that everything is going fine. What is fundamental for me in all this, is to make sure that the interests are in line, so that we can work together. The day the interests are out of line, there will obviously be conflicts, as we are not friends but partners. In this case, the investors with whom we have worked, in other words Iris Capital and E Source, were very good people with whom we have extraordinary good working relationship. But they don't have the same objective as us, meaning that they need to bring money back for their own investors, and us, we have a different perspective. Therefore, the relationships were successful, maybe also once more because, apart from the fact that they were good people, we were extremely careful to line up the interests, and with each step, to ensure that there would be no discrepancies. Therefore the company did not go go through extraordinary traumatic times when people had to be laid off, when potentially we could find ourselves in a divergent situation with the investors. Therefore not having experienced that, I did not lived through major conflicts with the investors, we almost agree, at the time of the sale, on the objectives, and from then on, ultimately, the path was as quiet as can be in a situation like that one. Then, the advice given to the creators of companies, first of all I would tend to be very careful about this because the paths are very often extremely different. Here we have experienced high technology company journey, where the commercial side has finally played a rather small role because we have sold the company before really comercialising the products. There is, without a doubt, one thing in common with most of the paths, it is the fact that still we start with an extraordinary idea and a very strong enthusiasm for this idea, it is extremely probable that our idea will change three, four, five times during the course of the journey. And for me, on of the very important things, which is very similar to a journey of discovery, basically I tend to think that the creation of company and a research project are paths very close together. It is the fact that it is necessary to be prepared to change ideas, at the same time, to have a vision, in other words which way we are heading, but at the same time being extraordinarily mobile, and adapt to the situation and change ideas. The second point which is in common with a journey of discovery, is the importance of people. It is necessary to recruit people are extremely good from the beginning. It is not like in a large company where anyone can be replaced by anyone. Our luck comes from the start, when we were a team of four people like Erwan Le Pennec, Jérôme Kalifa, Christophe Bernard and myself, who worked well together with completely different personalities, but the other three were really big talents, and the people we recruited after were really excellent in their field. And personally, it is one of the most important things, but once more, it is the same thing in a research team, you need excellent people right from the start because otherwise the company will head in the wrong direction. The standardisation will come later, at the time when the processus will be established, but at the beginning the excellence, is for me absolutely fundamental. And then as I mentioned, and it is part of the fun, changing our mind on a regular basis, is something which not only is possible, but I think, rather important in such a journey.
