IA PARC platform helps companies modernize their technologies, rethink their processes and transform their experiences
Denis Escudier - CEO
Over the last 10 years, data science has developed considerably thanks to the conditions finally met for its full expression. The processes of data digitization started in the 1980s and accelerated in the 1990s by the arrival of the Internet. It accelerate again in the 2000s by the growing conversion of companies to computer tools and the dematerialization of personal data for individuals which have generated unprecedented volumes of data. The problems posed by the cost and security of data storage were resolved thanks to data farms and the cloud at the end of the 1990s.
The performance of hardware and the miniaturization of chips made a quantum leap and allowed for increased storage at lower cost and increased computing capacity. Kryder’s law reminds us that the storage capacity of hard disks is doubled every 13 months while dividing their cost by 2. In 1981, the equivalent cost of storing an MP3 song (1MB) was $700. Twelve years later it was only $1 and after 2010 it was less than 10 cents per GB.
We moved from an economy of possession to one of rental and a multiplied annual production of data that acquires intrinsic value through its volume, variety and speed of refreshment. According to IBM, in 2019, 90% of the world’s data had been created in the previous two years. But these new economic challenges involving “21st century gold” are intimately linked to environmental and societal issues. Faced with this context, MomentTech began a CSR policy as soon as it was created in 2016, on the one hand by equitably associating employees with the company’s capital, and on the other hand, by looking for a local and innovative player in “green” data hosting.
Our first partner was Webaxys, the first true European, French and Norman “Green Datacenter”! Raw data becomes even more valuable and meaningful for organizations when it is analyzed, correlated and translated visually so as to easily read the trends of a company’s activity for example. This exponential increase in data volumes, translated into terms of “Big Data” as early as 1997 but democratized much later, has allowed the rise of “business intelligence” and “datavisualization” or “dataviz” from 2010 to today.
Our American friends have understood this way before us. They were the first to create the conditions for a high-performance and accessible cloud. Their lead is so strong and their hold on the European continent’s data so strong that GAFAMs now account for 70 to 75% of the data hosted in Europe according to an EESC report in 2021. They were the first to understand the “business” value of this stored and refreshed data. They were therefore naturally the first to offer tools for analyzing, exploiting and visualizing data, and more recently, the ability to orchestrate data and Machine Learning algorithms, and then the training of AI models in the cloud on neural networks for Deep Learning. This technological advance poses a crucial problem of sovereignty for the future of our continent if all the “gold” of the 21st century is held by 5 to 6 companies within a single country! The indispensable sovereignty of data implies the future of our economic independence.
This challenge was at the origin of our desire to offer an effective alternative to French and European companies: an autonomous Data Science platform, deployable in each company or accessible in a sovereign and secure cloud, allowing organizations to process their data and implement it in a secure manner. To go even further and sustainably transform organizations beyond dataviz, the “Big Data” phenomenon has enabled the promising development of artificial intelligence. Machine Learning has favored the implementation of more precise predictive analysis algorithms that have concrete applications in our daily lives (price prediction, spam detection, fraud detection, product recommendation…). Since the 2010s, Deep Learning has gone further and its use has contributed to the discovery of exoplanets, the development of new drugs, the detection of early diseases, the autonomous car….
Although most of us are more interested in the spectacular results of this scientific evolution, it nevertheless induces a delicate implementation and important upheavals in organizations: the training of new engineers, the sensitization of “data/IA citizens” within business teams, the retraining of impacted employees, the interaction of business and AI teams, and the creation of the tools necessary to exploit data and AI. Breaking the isolation of engineers and helping them accelerate the implementation of their artificial intelligence models are organizational and productivity issues. It is now a question of working in project teams on the same tool and being able to link up with the business and IT teams to finally succeed in industrializing projects that have often been compromised. This is why this “data science” and the implementation of the work of collecting, storing, cleaning, characterizing and analyzing data, as well as the search for the best algorithms and the implementation of neural networks, require a change in organization and method.
It is no longer a question of thinking about “POCs” but about projects and their industrialization, and to succeed in what Gartner denounced as a (temporary) failure of AI in 2020: “85% of AI projects fail”. When we created the IA PARC platform, the first sovereign and collaborative platform for artificial intelligence optimized for deep learning, we were aware of this organizational challenge. It considerably improves the productivity and technical working conditions of data scientists and AI engineers. They finally have all the tools they need in a single workspace to integrate and store their data, then process and analyze it and create their AI models.
It is now possible to share projects in teams, have them validated by business teams and put them into production with the interaction of IT teams. MomentTech is proud to have taken up the challenge of industrialization to take it even further. We are involved in research and standardization bodies to bring to fruition the trusted AI that will allow everyone to adopt more of the spectacular and exciting new applications of this science. Provided that we keep the idea, and this is the meaning of our work and our commitment, that the well-being of all and our natural environment are always placed at the heart of our ambitions.
Denis Escudier