Three major trends will dominate the information technology market for years: edge, cloud and data, says Luis Buezo, global director of Artificial Intelligence Services and Data Platform at Hewlett Packard Enterprise (HPE).
The order of the statement is revealing: the first – he insists – is the edge (extreme or edge), because there, on the far periphery of the datacenter, a huge volume of data is generated; it would make little sense to move them around, but they are essential for centralized decision making. “Modern data platforms – he points out – have to work in such a way that companies can ingest, store, process, catalog and exploit this distant data, preventing it from traveling through the network and sacrificing performance”.
Buezo is also responsible for the Global Center of Excellence in IT, which HPE launched in April in Madrid. To meet a growing demand, the firm has a battalion of specialists in AI models and data analytics. "My role is global, but in Spain we did a study some time ago whose conclusions indicate that practically all the clients consulted consider that AI is a strategic area for their companies, which is why they are going to take initiatives and invest in it."
As he relates, this branch of HPE services begins by identifying the AI use case that may be behind the need expressed by a client: what is he looking for? What data is available for analysis? or where are these data and what is their quality? It is up to the customer to decide if they are processed at the edge, moved to the data center, uploaded to a cloud, or hosted on a shared server. Each option will have effects on the performance of the applications that arise from the project.
Good, but how do you create an AI project and how does it differ from conventional ones in IT? “Initially, we verify the return on investment that the client expects, because many times, when you dive into the data, you discover that the expectation was not justified.” The following steps are classic: proof of value, design, commissioning and deployment.
He eloquently weaves together real use cases, though only a few of them can reveal the identity of the customer. An exception is the pharmaceutical company Novartis, with which HPE has developed a dengue outbreak prediction system: combining health and meteorological data, it identifies areas in which to take preventive measures. Medicine is at the forefront of his story: as a member of the European Gatekeeper consortium, HPE is contributing to the development of a predictive model of diabetes. Another example: with hospitals in various countries, it develops a radiology image analysis model, in addition to participating in another initiative for the distribution of blood platelets.
In industry, the second field of application, the obvious field is the automotive industry: “Our most interesting work is related to natural language processing applied to optimizing vehicle configuration. In the financial sector, we deal with two issues: the fight against fraud and improvements to the conversation interface”.
That the company decided to install its AI center of excellence in Spain responds to different criteria, according to its director: “Quality of the universities, availability of young talent, solid presence of HPE and successful previous experiences with other centers in the country. The fact that this practice of the corporation is directed from here will have helped, but the truth is that the team is multinational and multidisciplinary”.