Open to opportunities

Mateus Rocha

@mateusrocha

Message

I build production machine learning systems for risk, churn, monitoring, and scalable data platforms.

Brazil

Message

What I'm looking for

I'm looking to build and scale production machine learning systems, applying rigorous statistical methods, strong governance, and data engineering to support business decisions.

At BV, I optimize BigQuery and Databricks workloads, refactor PySpark pipelines, and consolidate production models, reducing infrastructure and maintenance costs by approximately US$18K per month.

I built a model-monitoring dashboard using statistical drift and anomaly detection, translating results into Power BI visualizations that resolved more than 15 Internal Audit and Model Validation findings. I also migrated 10+ legacy collection and billing models into standardized Python and SQL pipelines with reusable components and stronger governance.

Previously at Carpa Analytics, I developed and deployed risk, churn, and medical authorization classification models using large-scale production data from 7K+ dental providers. My Applied Mathematics Ph.D. research applies XGBoost-based forecasting and imputation to meteorology, including custom evaluation metrics and a specialized pipeline for imbalanced regression.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Oct 2025 - Present (10 months)

Optimized BigQuery and Databricks workloads, refactored PySpark pipelines, and consolidated redundant models, reducing infrastructure costs by approximately US$18K per month. Built a model-monitoring dashboard with statistical drift and anomaly detection, resolving 15+ audit findings, and migrated 10+ legacy models to standardized Python and SQL pipelines.

Python SQL PySpark Databricks BigQuery Power BI Model Monitoring Anomaly Detection MLOps Model Governance

Data Scientist

Carpa Analytics

Oct 2023 - Oct 2025 (2 years)

Developed and deployed risk and churn prediction models for customer renewal and provider retention, analyzing production data from 7K+ dental providers. Built a classification model for medical request authorization and enhanced MLOps reliability by defining maintenance, monitoring, and deployment standards.

Python SQL Machine Learning Risk Modeling Churn Modeling Classification MLOps Monitoring

Education

Degrees, certifications, and relevant coursework

University of Campinas (UNICAMP)

Doctor of Philosophy, Applied Mathematics

2022 -

Ph.D. in Applied Mathematics with research on XGBoost-based time series forecasting and imputation applied to meteorology. Developed a novel Priority Queue Method and custom evaluation metrics for an imbalanced regression problem.