Matheus Caetano

• Directed the migration of the company's entire file system into a scalable Google Drive architecture, eliminating duplicate file versions and accelerating project delivery cadence • Developed a Python desktop app that reduced repetitive administrative processes from multi-step workflows to a single ~3-minute task • Automated ingestion and transformation of external data sources by building ETL pipelines in Power Query, turning a daily manual task into a single monthly process

• Engineered an ETL pipeline, consolidating 1.64M+ rows of multi-sensor time-series data from CSV files, including cleaning and feature engineering • Applied dimensionality reduction (PCA, UMAP, t-SNE) and unsupervised clustering (K-means, HCD, DBSCAN) to analyze 30-dimension matrices, retaining over 90% of data variance • Translated a physical sensor-selection problem into a mathematical optimization task, selecting the optimal 4-sensor combination with the highest orthogonality