Esteban Caro

Developed ML models to predict customer recovery and delinquency, enabling proactive risk mitigation. Built early warning systems to identify risk transitions and behavioral deterioration. Designed backtesting and validation frameworks to ensure model robustness. Built LLM-as-a-judge evaluation pipelines for conversational AI on AWS (SageMaker & Bedrock), defining reliability metrics.

-Designed dashboards in Amazon QuickSight to facilitate monitoring and alerting of collection management.
-Proposed, designed, developed, and executed ETL pipelines for portfolio data to identify behavioral patterns, detect alerts, and suggest adjustments in the collection strategy.

Updated the imputation methodology for a key variable in the credit risk evaluation of customers.
Conducted descriptive and exploratory analysis of large datasets, identifying trends and key variables to understand users' financial profiles.
Developed a machine learning model to estimate the maximum financial quota assignable to clients based on their risk profile

Engaged in a course on Databricks: Delta Lake Architecture, learning about data lake architecture and its applications in big data processing.

Pursuing AWS Cloud Practitioner certification, focusing on cloud computing fundamentals and AWS services.

Studied Industrial Engineering with a focus on data science applications. Developed skills in data analysis, machine learning, and statistical methodologies to solve engineering problems.

Completed a specialization in Artificial Intelligence for Engineering, focusing on machine learning and its applications in engineering problems.