Looking for a job

Juan Osorio

@juanosorio1

Data Engineer at Krones AG who built production pipelines processing 2+ TB/week and migrated them to AWS MWAA.

At Krones AG, I built Python and SQL workflows that integrated SharePoint and SAP exports for operational reporting. I also created Power BI dashboards used by European and Central Asian operations.

At Wizeline, I owned an Airflow and Spark production pipeline processing 2+ TB/week, with a focus on recovery and scheduled delivery. I coordinated its migration to AWS MWAA without interrupting downstream delivery.

At MO Technologies, I trained credit-default models and implemented lending-policy rules on top of model predictions. Evaluating decision thresholds helped reduce the default rate by up to 15% relative.

At Leal SAS, I developed a Python pipeline for engagement monitoring and customer campaigns; A/B tests attributed a 10% increase in average transaction value to the campaigns. I also reduced peak database CPU utilization by 30% through changes to job scheduling and data retrieval.

Experience

Work history, roles, and key accomplishments

Data Engineer

Krones AG

Jun 2024 - Dec 2025 (1 year 6 months)

Built Python/SQL ETL workflows to integrate SharePoint and SAP exports, standardizing schemas and business fields for operational reports. Established validation rules for curated dashboard datasets, reducing errors in Power BI reports.

Python SQL ETL SharePoint SAP Power BI Git Documentation

Software Engineer III – Data Engineering

Wizeline

Sep 2021 - Aug 2023 (1 year 11 months)

Automated application and AWS infrastructure releases across environments with GitHub Actions and AWS CDK. Processed multi-region Prometheus logs with AWS Glue/PySpark for analysis in Athena and QuickSight, helping infrastructure teams identify idle servers and target cost reductions.

Python SQL Spark PySpark Airflow AWS AWS CDK GitHub Actions AWS Glue Athena QuickSight Prometheus AWS MWAA

Data Scientist II

MO Technologies

Aug 2020 - Sep 2021 (1 year 1 month)

Trained credit-default models on client history with Python, scikit-learn, XGBoost, and PyTorch. Reduced the default rate by up to 15% relative after evaluating class separation with distribution analysis and KS tests and selecting decision thresholds.

Python scikit learn XGBoost PyTorch S3 Athena Pandas NumPy Matplotlib Model Monitoring KS Test

Jr. Data Scientist

Leal SAS

Mar 2019 - Mar 2020 (1 year)

Combined customer and transaction data with Python/pandas and PostgreSQL/MySQL joins, aggregations, and reusable views to calculate lifetime ROI and customer-behavior metrics for recurring analysis. Cut peak database CPU utilization by 30% through changes to PostgreSQL/MySQL job scheduling and data-retrieval patterns.

Python Pandas PostgreSQL MySQL A B Testing Data Analysis