I've built and optimized end-to-end machine learning pipelines for banking, financial services, and Coca-Cola FEMSA projects in Mexico City.
At Capgemini for Banamex, I reduced a legacy ML pipeline from 2,500 to under 800 lines using Keras, scikit-learn, pandas, NumPy, and MLflow. I also developed NLP sentiment analysis and anomaly detection models for brand reputation monitoring and large-scale banking migration processes.
More recently, I developed a GCP Vertex AI sentiment analysis pilot for Coca-Cola FEMSA, connecting Twitter/X data to BigQuery and LLM-based scoring while monitoring production models and databases through Azure Portal. I focus on production metrics such as Recall and PR-AUC, with hands-on experience across Python, cloud platforms, NLP, deep learning, ETL, and MLOps.
