IT Analyst
I followed 3 main activities concerning:
Credit Granting Platform:
Developed software for managing various forms of credit (personal loans, credit lines, mortgages, guarantees) for both private and corporate clients, integrating external systems via REST APIs and banking’s legacy systems.
Credit Risk Monitoring:
Implemented solutions for monitoring and managing credit risk when extending additional credit to existing clients, using predictive analytics and risk evaluation models.
Dashboarding for Collateral Value:
Created interactive dashboards to visualize in real-time the value of assets pledged by clients, optimizing decision-making in the credit process.
The first point was certainly the most challenging and related to the platform that regulates the granting of credit to both private and corporate customers. Credit could take the most varied forms: personal loans, bank overdrafts, mortgages, etc. All forms of credit were managed by the same platform with different dynamics.
Data Engineer for ML Applications
I worked with a small team to bring the first Machine Learning models adopted by the bank into production, starting from scratch. The activities, in a very synthetic way, are described below:
Data Integration: Management of the ingestion and transformation of data from different sources into the datalake, ensuring a continuous and centralized flow of data for subsequent use in Machine Learning models.
ETL Pipeline: Development and management of ETL (Extract, Transform, Load) pipelines to transform and transfer data from the datalake to the Dataiku platform, optimizing the quality and structure of the data for modeling.
ML Development and Implementation: Support for the development and testing of Machine Learning models within Dataiku, ensuring that the pipelines supported the needs of model update and accuracy.
Monthly Automation and Update: Implementation of automated processes for the monthly ingestion of new data, with continuous validation of the quality to ensure constant and accurate performance of the models in production.
