Making a travel agent do less work: context selection with JEV ↗
A same-model comparison of a legacy travel agent and JEV: context selection, additive tool loading, measured latency, cost, acceptance and cache behavior.
AI/ML Engineer & Data Scientist
I’m a physicist working at the intersection of machine learning, AI engineering and software development. I turn open-ended problems into systems that can be evaluated, deployed and maintained.
At Intelequia, I lead the Data Science team across applied ML, intelligent agents and MLOps. My work connects technical direction with hands-on delivery, from choosing the right approach to deploying models and monitoring them in production.
Technical leadership and hands-on delivery across applied ML, GenAI and production engineering. I work from problem definition through deployment, with evaluation, governance and monitoring built into the process.
Lead the Data Science team and shape the technical approach across modeling, Azure, Databricks and MLOps. Own practices for model versioning, promotion across environments and production monitoring. Build Azure AI Foundry agents grounded in client data, with evaluation gates before release.
Moved ML from ad hoc runs to repeatable training and inference workflows with Docker, Kubernetes, ArgoCD and MLflow. Delivered DISA’s first GenAI use cases end to end, including RAG and sentiment analysis. Improved route solvers and led a joint anomaly-detection project with the University of La Laguna.
Built the data pipeline that turned large volumes of marine audio into research-ready datasets. Developed sound classification and event detection models to support analysis of the environmental impact of anthropogenic underwater noise.
Maintained data integrity in the tourism data warehouse and translated its data into statistical analyses and Power BI reporting for the research team.
From physics to data science and astrophysics. A background in understanding complex systems, testing assumptions and working with uncertainty.
Universidad Internacional de Valencia (VIU)
Completed June 2025
Universidad Internacional de Valencia (VIU)
Completed June 2023
Universidad de La Laguna (ULL)
Completed September 2021
I’m building Travel Ruter as a product people can use to plan their own trips. Routes, saved places, budgets, notes and editable guides, together in one workspace.
I own the product from its interface to its backend and AI assistant. The assistant helps people work with their own saved places, notes and itinerary.

A same-model comparison of a legacy travel agent and JEV: context selection, additive tool loading, measured latency, cost, acceptance and cache behavior.
How Mnemosyne combines hybrid vector search, cross-encoder re-ranking and a Claude ReAct agent to detect duplicate bug bounty reports.
Professional credentials in ML operations and data analytics.