BACK TO INDUSTRY INSIDER

Voice Recognition: From R&D to Start-Up Creation

Ergonomic Technology at Your Service

Voice Recognition: From R&D to Start-Up Creation

Francis Charpentier, chief scientific officer (CSO) and founder of France Telecom's spin-off Telisma, followed his passion for voice and sounds to specialise in speech software and service. He is based in Lannion, Brittany, France.

Having trained as an engineer at one of France's most prestigious schools, École Polytechnique, Francis Charpentier specialised in telecommunications through 1 year at École Supérieure de Télécommunication in Paris. With his excellent academic profile, the R&D branch of France Telecom snapped him up when he completed his studies in 1981. He took the opportunity of working in R&D to do his PhD, the aim of which was to "make machines speak." "What attracted me towards [voice synthesis] was a technology that combines computer science and mathematics together with linguistics," says Charpentier. He enjoys working in a multidisciplinary field with concrete applications.

Back in the '80s, France Telecom typically offered a job for life, but Charpentier knew that staying put was not the way to advance his career. So after completing his PhD in 1989, he decided it was time for a change. He got a job at the Cap Gemini consultancy working on a European R&D project called Sundial, financed by the EC's 3rd Framework Programme. There he worked on voice recognition with application in enabling human-machine dialogue--a natural step after his PhD. "The work consisted of integrating intelligence and ergonomy in the dialogue with the machine," he says. A typical application of this work is the automated airline telephone booking service. "This experience gave me a taste for working at an international level," he says.

When Cap Gemini decided that this type of research was no longer their priority, Charpentier went back to France Telecom R&D. This time he had the opportunity to work on technology transfer projects, negotiating licensing agreements with companies that wanted to exploit France Telecom's patented technology. "This was a first step towards [the business side of] industry," he explains.

In the '90s, the voice technology field became much more competitive as US start-ups were founded to exploit voice recognition systems commercially. "Seeing the competition gave me the idea to set up my own company," he explains. "I knew our technology was good and had reached a level of industrial maturity," he adds.

The concept for Telisma, based on speech software, was put together in July 1999. And the company was officially created in September 2000--with France Telecom retaining a 20% share. For the first 6 months, Charpentier was involved in all aspects of the business. When investors stepped in, they brought in a CEO and assigned more focused roles to the team. As a result Charpentier became CSO, and he started doing less research and more development to meet clients' demands.

To people who would be interested in creating their own start-up, his advice would be: "You have to take a calculated risk." Clearly, the support of a company like France Telecom helped tremendously in the beginning. But the drawback of being in the commercial world is that "customers want perfect solution at low cost, whereas such technology is expensive because of the amount of research involved," he warns.

And for those who dream of a career in the telecoms sector, he reckons that having a job for life is not the solution. "What matters most is mobility," says Charpentier. Nothing replaces the experience gained from several jobs. And he believes that holding multiple positions in the industry teaches people to be flexible.

Ergonomic Technology at Your Service

Alex Hum has brought the advanced knowledge he acquired in his native Singapore to Europe and is working at the interface of technology and human interaction as a senior R&D manager with Orange.


Alex Hum

"I did my PhD working in a very new area which was the focus of the industry," explains Alex Hum, pictured at right. The project, which was based at the National University of Singapore (NUS), but supervised from the University of Waterloo in Canada, focussed on system design for advanced wireless communication.

He finished it in 1996 and was quickly recruited by the Centre for Wireless Communication, a NUS spin-off formed to exploit the commercial potential of the university's R&D. There he worked on the development of a software radio architecture with SGS Thomson, which is used in future 3rd generation (3G) handsets.

At the same time, Hum had an opportunity to teach an MSc course in radio and wireless communication system design at the NUS. "This was a very innovative, hands-on course to bring students up to speed with the latest wireless technologies," he explains. The course was so innovative that it attracted a lot of attention from foreign universities, such as the University of California, Los Angeles..

Such exposure, and publications in peer-reviewed journals, also attracted the attention of prospective employers. In 1999, he was headhunted and received two offers: the first from a national research company in the US, and the other from the Starlab research laboratory in Brussels, which was created with the aim of competing with the Media Lab at Massachusetts Institute of Technology (MIT). Unfortunately Starlab went bankrupt in 2001.

"I decided to take the European offer," says Hum. "Europe was attractive because they are more advanced on this type of technology, and also because of the culture," he adds. He became a research manager for several areas including mobile communication and content management. He also worked on the development of applications of wearable technology--invisible technology that is designed to enhance simple and intuitive human interaction. "This was like having the best of both worlds," he says: having the opportunity to develop sexy technology while making access to that technology easier.

To make the industrial consortium on wearable technology happen, he had to turn to industry to get funding, managing to gather 20 sponsors from the technology and fashion industries. Although the research initially seemed rather blue-sky, it required the development of a lot of intermediate technologies that could be patented and that were immediately applicable, such as an innovative wireless communication infrastructure. This idea won him the International Innovation Prize in Germany.

In 2001, Hum moved to London, when he was invited to become chief technology officer at the Starlab spin-off, Yeahlab. Yeahlab was created to deal with mobile content creation and management and products such as SMS, MMS, wap, and picture messaging to meet the growing consumer demands of Generation Y--the trend-sensitive generation of technology-savvy teenagers.

Developing these new projects involved working closely with Orange, which made him an attractive offer he could not refuse. "It's my dream to work in Orange," he says. His new role is to introduce new ideas and innovations to the group. "They appreciate my entrepreneurial and technical background," says Hum. His previous experience, as chief scientist and research manager of an industrial consortium, as well as in a think tank like Starlab, helped a lot in securing him the position. Orange realised he has the ability to mix innovative ideas with a business sense. In other words, he already had experience making money from innovation and spotting new ideas that could be commercially viable.

At the same time, Hum had the satisfaction of seeing some of the ideas he worked on back in Brussels brought to the market with Orange's launch of the Wearaphone at Christmas last year, the company's first wearable product. This is the real attraction of working in telecoms for Hum: "to have the opportunity to help influence technology and the future of telecommunications and services." He believes that there is a "need to be far thinking but yet understand customer needs" in order to provide better products for the consumer. And he admits that he likes testing new devices and having a direct link with consumers.

On the recruitment side, "we need people who think out of the box and are not afraid of being creative," notes Hum. And "to get acceptance of your ideas, you will need to be supported by a business model," he adds. His advice to someone who is interested in developing a career in industry from a research background is to "get as much exposure as possible to the industry" in order to understand the criteria of entrepreneurship and business models. In that way, he says, it will be possible to see how the research can benefit consumers.

Sabine Louët is a freelance writer based in Dublin.