
Data Analyst Intern
EY (Ernst & Young)
·Jul 2025 – Sep 2025 (3 mos)
1 / Challenge Context
The Alcobev domain struggled with unpredictable cash flow patterns due to siloed data sources and manual KPI tracking. The lack of a unified predictive model led to inefficiencies and delayed decision-making across multiple business units.
2 / Strategic Implementation
Engineered a robust forecasting system leveraging FastAPI for the backend and advanced ML pipelines powered by XGBoost and LightGBM. Standardized data preprocessing steps and built a flexible architecture to allow cross-domain validation and rapid deployment of new models.
3 / Quantified Impact
Automated 13+ critical KPIs and achieved an R² accuracy of 0.87. The universal engine reached 95% interoperability, successfully validating across 4+ business domains and significantly streamlining predictive operations.