Today, banks are looking for their own ways to implement tasks associated with the digital transformation of business. In his interview Maxim Kondratenko, member of the board of VTB Bank, Ph.D. in Economics, talks about how this process is taking place in VTB and what changes it makes to the creation of new products demanded by customers, how it affects the increase in the share of services in digital channels, the development of data analysis tools and sales support, as well as increase in labor productivity and helps reduce the cost of operations. He shares his impressions of the project, carried out jointly with Neoflex, as a result of which the bank received the Data Lake based on Hadoop technologies to work with data fr om external sources.

Fragment of the interview

CNews: What are the advantages of using Data Lake when compared to a traditional data warehouse?

Maxim Kondratenko: There is an instrument for each task. Data Lake is effective where it comes to a large and—above all—external volume of heterogeneous information that needs to be structured and processed quickly. In addition, Data Lake is the ability to quickly test hypotheses and make changes to Agile principles. Traditional data warehouse is less suitable for such experiments.

We called working in Data Lake with new external sources of information, the results of which are applicable in various risk management processes, Kantor project. The way we will move is to increase the number of elements of the credit procedure, in which the information obtained in this way is used. Monitoring of credit risk factors is one of these elements and the first pilot, the benefit from the implementation of which we must see quite quickly.

CNews: How did the name of this project appear? What is the meaning of it?

Maxim Kondratenko: This is an abbreviation — Brand New Technique of Risk Management — with a deep inner meaning. In general, people who are associated with technology, like to come up with original names. In our case, it was consonant with the name of the great German mathematician, the creator of set theory.

CNews: The Kantor project turned out to be significant in terms of scale. And, surely, VTB has an IT provider that helps implement it? By what criteria did you choose it?

Maxim Kondratenko: VTB Group has been acquainted with Neoflex for a long time, since the creation of corporate data warehouse for VTB Factoring and working with VTB24: development of integration solutions and transactional services, carrying out speed-torque and automated functionality testing, support for various IT systems and implementation of observation of work. Now we have a positive experience with Data Lake, big data and machine learning. From the point of view of a project-based approach, the company demonstrates good skills and a culture of development in the ideology of DevOps. This facilitated the implementation of a pilot project from infrastructure deployment to the receipt of the first calculations for three months.

CNews: What other practical tasks can be solved using Data Lake?

Maxim Kondratenko: Data Lake combined with the stack of Hadoop technologies gives users almost unlimited possibilities in the use of Data Mining tools. Therefore, sandboxes of data were deployed within the data warehouse, wh ere Data Science specialists conduct various experiments with models not only directly in the interest of risk management, but also to analyze the client profile in the framework of CRM, to search for behavioral anomalies, to forecast account balances, to prepare cross- , upsale-offers and many others. Thus, the global goal is to move from descriptive analytics to models based on machine learning methods, predicting events, parameters and providing recommendations.

Full version of the interview is available on the website.

Source: CNews

Press service

Have questions or need more information? Please call the press office on the number
495 984-25-13

Or send an email to marketing@neoflex.ru

We are happy to help!

Subscribe to newsletter

Be aware of fresh developments in the «Neoflex»


Share

Back to publications