Odei Hijarrubia

• Built the ingestion side of a public-tender intelligence platform for Mahou San Miguel on Palantir Foundry: real-time RSS capture, LLM extraction of ~30 fields from structurally inconsistent PDFs, and the ontology that unified them — ~300,000 historical tenders in the initial backfill.
• Designed ETL and machine-learning workflows in Python and Apache Airflow across client engagements, from DBSCAN clustering for entity resolution to LLM-driven classification pipelines.

Python · SQL · Apache Airflow · Palantir Foundry · Microsoft Fabric · LLMs · DBSCAN

• Engineered a predictive-maintenance model for the A350 pneumatic system from multi-sensor time-series data, with feature engineering targeted at forecasting system leaks.
• Defined automated model-testing strategies so the work could stand up to a real industrial maintenance setting rather than a held-out split.

Python · Pandas · Scikit-learn · SQL · Spark · Palantir Foundry

• Built full-stack features for the Government of Navarra’s surgical-interventions management application in C# and .NET over SQL Server, in an Agile team of five.

C# · .NET · SQL Server · Git · Scrum