Pedro Luis A.

Most business teams I've worked with have the data — it's just buried in SAP exports, manual Excel files, or reports nobody reads. My job is to turn that mess into something people actually open and use to decide. I've spent 12 years doing this inside John Deere and Volvo spare parts operations — not as a consultant who comes in with a framework, but as the person who built the systems from scratch. I architected the SAP ↔ John Deere RPM integration that runs purchase orders automatically. I built the forecasting models (XGBoost, CatBoost, LightGBM) for my PhD. And right now I'm using Claude AI to automate decisions that used to take days of manual work. The way I work is simple: before touching any data, I ask "what questions are we trying to answer?" That one habit saves weeks of building the wrong thing. What I build: Decision tools — Streamlit apps · AI agents · automated reporting Data infrastructure — SAP · SQL · Snowflake · ETL pipelines BI Dashboards — Power BI · Tableau · inventory & commercial KPIs ML & Forecasting — XGBoost · LightGBM · demand modeling · Python Inventory Management and Replenishment Data Driven Systems Industrial Engineer · MBA · PhD in ML-based inventory optimization. Bilingual EN/ES.