At Minsait, I develop end-to-end AI solutions for public administration, infrastructure, and heavy industry clients. For a regional administration, I benchmarked open-source LLMs and developed fine-tuning and a retraining lifecycle for a model specialized in a co-official language.
For an infrastructure project, I developed the backend and frontend of an application that uses generative AI to validate tender-document criteria. Its map-reduce pipeline classifies and scores documents with LLMs.
I also worked on an incident-classification system using a neural network, SVM, and nearest-neighbor information retrieval, comparing approaches to maximize accuracy with limited resources.
Since 2024, I’ve collaborated with ISEO Energía on SYGEO, a cloud platform on Azure that brings together energy-supply data to help companies make decisions about electricity spending. My AI degree at Universidad Rey Juan Carlos included a final project on few-shot energy-consumption forecasting over time series.

