CI&T, Databricks, and a World Leader in Consumer Credit Reporting

Improving efficiency and scale in data management
Know more
The project, developed in partnership with the brand, saved US$ 4M in 18 months, while also reducing customer implementation time by 75%
Know more about our solutions

The challenge

The partnership between CI&T and the credit analysis company began in 2017, aiming to build the Positive Register. 

Individuals' credit scores in Brazil have always been calculated based on unpaid bills, or default. In turn, the Positive Register emerged from a change in legislation aimed at contributing to the score of all bills paid on time by the consumer. With a much larger volume of data to process, the company had a technological challenge to build a robust structure to support this new reality.

Solutions

Both companies organized the work on three fronts.
On these fronts, we took a transversal look at our client's data architecture, from enabling the Positive Register to recreating the previous legacy infrastructure.
01.

The construction of the Positive Register

02.

Data Management Platform: Migration from a solution with high licensing costs and low scalability to a more modern, economical, and faster solution delivery model

03.

Analysis Platform: Development of technology to democratize the company's credit score for its customers. This will give large banks autonomy in creating data models based on their demands without centralization

Positive Record: Using Databricks for Fast Learning

A hybrid infrastructure was chosen by the company to increase the computational power of its clusters. The processing was done in two places: on-premises infrastructure and cloud infrastructure.

AWS scaled the infrastructure to process data and Databricks – a partner that supports complex code execution – removed friction in business areas.

Creating a robust data processing cluster and an environment for formulating business rules and codes is relatively easy. The data area has room to experiment without immediate concern about the data structure. With Databricks, engineering teams build experiments with agility for testing and learning, just as business teams use it to create machine learning models.

Thus, different solutions, credit models, etc., could be tested quickly. CI&T acted as a connector between the various areas for Databricks adoption. More than 120 engineers worked on the project.

Want to learn more about our work?

Contact Us