Though data is helping us to grow multifold, a lot of time goes into data cleaning and data segmentation. Structured and unstructured data is to be segmented and then the datasets so created are used for machine learning and deep learning algorithms. Ingedata proposes industrialized services to structure and prepare datasets that are used to train computer vision and NLP models. Its differentiation is that it implements client-specific production lines for dataset annotation. Teams of annotators are based on its own production sites to keep high control on production, annotators are recruited to meet domain expertise requirements, and Ingedata’s support departments (methods, project management, training, HR) create structured and repeatable processes.
This positioning proves to be valuable to highly technical projects, which involve challenges such as the access to specific talents, the management of complex processes in ambiguous environments, the compliance with GDPR and confidentiality requirements, and the ability to scale production.
Purpose Since the Creation
Prior to being an “AI company”, Ingedata’s value resides in the ability to manage large production pools. Since the creation of the company (12 years ago), its core activity has been to set production lines to enrich and structure large databases. This is what the company now applies to industrialize and scale its Client’s computer vision and NLP projects.
In addition to that, Ingedata is a truly internationalized company and is an important recruiter in developing countries. Its mission is to integrate into these countries’ economic ecosystems, provide job opportunities to the locals, and contribute to the countries’ involvement in highly innovative sectors such as AI.
The Visionary Leader
Pierre Le Barbier, CEO of the company, implements his vision to provide top-quality annotation services to complex projects. Delivering annotation for complex project implies deep understanding of the ecosystem, creation of a relevant business model, sourcing and training of reliable and meticulous annotators, definition of highly controlled production and quality inspection processes, strong IT skills to refine the annotation tools, and project management best practices. Pierre’s role in the company is to constantly align this range of actions towards the same goal.
Delivery of Cutting-Edge AI Models
Ingedata’s role in the AI industry is to catalyze the delivery of highly performing models. Data scientists need high-quality datasets to train their models and should not spend all their time cleaning data. When a client needs to get a model out of the lab and to convert it into a scalable product, the usual barrier is to manage data production. It is not a data scientist’s job. Ingedata becomes a key partner for this scope of the projects.
Newness to Intelligent Applications
It only takes one to read the news to understand that these technologies are at the core of innovative projects today. What’s great at Ingedata, is that its multisector approach allows to observe this change from the field. It is a cross-industry disruption that applies to all parts of the business. The company only sees the tip of the iceberg and new ways of applying AI. Ingedata faces new applications and uses almost every day.
Key Player of Innovation
What’s very innovative at Ingedata is that, while it is at the heart of the AI ecosystem, the company focuses on data rather than models. It is a unique approach that aims at supporting all the players of the industry – who usually focus on the algorithms. Its take is that, at the end of the day, data is the key to the success of AI projects. Making data usable and relevant is the success factor for such projects. It is an area where R&D focuses less energy as it is more about industrializing a process than about making new technical discoveries. Ingedata believes that it can make a big difference, by providing strong understanding on data matters, while its clients have the expertise in modelling.
In this context, Ingedata has partnered up with AI agencies, AI startups and researchers to make new projects scalable and competitive. Ingedata plays a role in proving to the end-clients that a project is scalable, in adding value to startups that need structured data to raise funds, and in adapting research findings to business-oriented projects.
High Value Added Through AI
Jean believes at the end of the day, AI opens up a set of new tools to further add value to businesses. Today, a lot of uncertainty still lies in AI projects and most applications are developed by R&D teams. The company sees the future of AI as a commodity that is adopted by business-centered services rather than R&D teams. AI is not becoming technically much easier in the future, but certainly more framed and controlled – just like a project to be managed among others.
Experimental Approach Against Challenges
The main challenge the company faces is that it provides industrialized services for data annotation, while AI projects still work a lot with a try and fail approach. It takes a lot of time to frame a project, get it out of the lab, and start a production line for data annotation. The company, however, sees the market evolving towards more structured needs and focus a lot of energy on evangelizing how to upscale projects.
Achievements in Competitive Atmosphere
One of Ingedata clients told the company that it had become a strategic partner to them and that the company had been their leverage for competitiveness and scalability. This is exactly what Ingedata aims at, and the company is proud to be a catalyzer to its Clients’ development.
The Growth of Algorithms for a Better Future
Jean feels a concern could be that algorithms become so accurate, so broad and so autonomous that there is no more need for data annotation in the context of supervised learning. “We, however, do not believe in one algorithm for all. We believe that a higher number of very specific algorithms will be developed to do extremely focused predictions and detections. In this sense, the way we see the company in the future ahead is working on even more narrow and technical applications, he added.