January/2022 - August/2022
As a Data Scientist at Involves, I worked on developing innovative solutions for the retail sector, focusing on data analysis and predictive modeling to optimize operations at points of sale.
Developed a proprietary mathematical model to predict expired products in retail stores, without relying on pre-built machine learning libraries.
Designed a probability simulation framework that estimated expiration risks for thousands of SKUs, overcoming computational constraints through parallel processing.
Helped double the company's client base within two months, as the solution became a key marketing focus.
Learned how to design custom statistical models from scratch and optimize computational performance in large-scale retail applications.
Tools: Python, computational simulations, statistical analysis.
Improved the accuracy of an existing out-of-stock detection model by replacing an unsupervised approach with an XGBoost-based model, increasing precision from 30% to 70%.
Integrated turnover rate and historical sales data, improving product replenishment efficiency.
Learned how to transition from rule-based heuristics to supervised learning, significantly improving model performance.
Tools: Python, XGBoost, SQL.
Created detailed Power BI dashboards, displaying key insights such as model accuracy, engagement metrics, and financial impact per store.
Automated data collection from multiple sources, improving visibility into retail operations.
Tools: Power BI, SQL.