● Developed and deployed Machine Learning models for sales personalization algorithms using Python, Azure, Databricks, MLFlow, and Spark, optimizing customer engagement for the Andean region at the world's largest brewing company. Orchestrated end-to-end Databricks workflows. ● Developed a time-series forecasting model, achieving a 25% increase in volume conversion and 3.9% increase in revenue, using Facebook’s Prophet and LightGBM to improve sales’ personalization in Paraguay and Bolivia by making forecasts for each brand pack and client. ● Re-engineered, deployed and monitored a complex collaborative filtering recommendation system using sci-kit learn, MLlib, databricks workflows, MLFlow, and spark for distributors/points of purchase for the world’s largest brewing company. Implemented MLflow model registry, health metrics monitoring and CI/CD, and deployed retraining jobs; obtained 60% accuracy initially, refactored and monitored, and after implementation achieved 30% expansion in the customer brand pack portfolio in the Andean region. ● Managed large-scale Machine Learning pipelines handling memory and time complexity, and monitoring periodic results to process millions of records, ensuring scalability and efficiency in sales’ personalization across the Andean region. ● Designed and executed A/B tests to evaluate ML solutions' impact on sales and engagement, applying statistical analysis and experimental design to drive data-informed decision-making. ● Collaborated with business stakeholders and domain experts to identify key sales challenges, and opportunities translating business needs and KPIs into functioning and scalable ML solutions. ● Designed and executed A/B tests to evaluate ML solutions' impact on sales and engagement, applying statistical analysis and experimental design to drive data-informed decision-making.

● Developed automation services for Paraguay’s largest quarry production and leading civil engineering company using Python, C, SQL, and embedded Linux, enhancing data availability for data-driven decision-making. ● Designed and implemented OLAP tables from transactional data using PostgreSQL, psycopg2, and Pandas, enabling efficient data readiness and reporting for production insights. ● Worked closely with engineering teams and field operators to ensure seamless automation and data acquisition.

● Developed a C++ implementation of ByteTracker (a renowned Multiple Object Tracker) for OpenCV by using Kalman filters and implementing a multipurpose linear assignment solver (Hungarian method) for robust object tracking; integrated IoU (Intersection over Union) metrics for precise object association and motion prediction. ● Contributed open-source code to the OpenCV repository, ensuring high coding standards in a CI/CD environment by using C++ interfaces, OOP principles, GT ests for unit tests, Python bindings and CMake to develop scalable cross-platform code for seamless integration across various environments. ● Managed code integration with Git, GitHub Actions, and Docker, ensuring reliability and seamless deployment.