Currently contributing to a large-scale Machine Learning project for Audience Segmentation.- Streamlined repeated code into functions improving maintainability and readability.- Expanded on feature engineering utilizing the DataMart I designed.- Modified the code to accept multiple algorithms instead of only XGBoost for hyperparameter optimization and best model selection. This improved ROC-AUC at least 5%, depending on the audience.Leading the development of an Audience Builder… Show more Currently contributing to a large-scale Machine Learning project for Audience Segmentation.- Streamlined repeated code into functions improving maintainability and readability.- Expanded on feature engineering utilizing the DataMart I designed.- Modified the code to accept multiple algorithms instead of only XGBoost for hyperparameter optimization and best model selection. This improved ROC-AUC at least 5%, depending on the audience.Leading the development of an Audience Builder tool for the Marketing team, projected to reduce manual audience creation time by 80% and improve campaign targeting efficiency. Show less

Designed the logic for Marketing`s first DataMart and oversaw its implementation.- Unified all of Despegar`s scopes (transactions, traffic, engagement) with over 85 tables into a client-centric model of 34 tables for Data Analysis, Machine Learning and KPI calculation, reducing query execution time by 90%.Structured a Python script for LTV calculation for any audience given a SQL query, allowing SEM investment per audience taking into account future spending.

Substitute teacher for three secondary school courses, where I explained complex math topics in simple and engaging ways to students