Develop machine learning model to assess creditworthiness of new borrowers, leveraging alternative data such as In-App User Behavior, SMS, Mobile Apps and Phonebook data from user's cellphone. Successfully deployed the model to production for real-time scoring, which has already served more than 2 million users in the Philippines and is projected to improve profits by 40%. Foster strong cross-functional collaborations with Data/ML engineering and business teams to ensure strategic alignment. Enable model training with a single command and efficient tracking of experiments, reducing time spent on model training by 5x.

Build production-quality data flow using software engineering best practices like version control, testing, CI/CD (dbt, SQL, AWS). Collaborate with a variety of business stakeholders to lead projects to completion. Balance multiple projects to meet deadlines.

Own the model lifecycle from ideation, training, evaluation to model monitoring of production-level machine learning models. Increase customer engagement on the alerts of a model by 20% by reducing monthly alerts by 3x from 300 to 100, and maintaining catch rate. Automate model monitoring using Airflow pipelines to label events, compare predictions with events and calculate metrics. Productionize ad-hoc data analysis model on the cloud using Data Modeling, Docker, Jenkins, which was released as a new product offering Uptake Compass.

Build R packages to REST APIs to automate complex requests to the World Bank Data Catalog. Build newspaper-quality data visualizations following the principles of good data visualization. Ranked in the top 5% among 2500+ candidates in the "Predicting Poverty" competition hosted by World Bank on DrivenData.