Freelancer profile translated to English.

Description

AI Engineer with experience building machine learning and artificial intelligence systems in production. I have developed conversational agents with Vertex AI and LangGraph for public administrations, ETL pipelines processing 3 million records daily, and classification models with 92% accuracy on real datasets.

I primarily work with the GCP stack (Vertex AI, Cloud Run, Cloud Build, Secret Manager), Python, FastAPI, LangChain, and Docker/Kubernetes. I have experience in both data engineering and in modeling and integrating LLMs into production systems.

If you need someone who delivers clean, well-documented, and production-ready code, let's talk.

Languages

Spanish
Native or bilingual
English
Fluent

Workplace preferences

Remote only

Primarily works remotely

Atos
AI/ML Engineer
September 2025 - Today (9 months)
Tenerife, Santa Cruz de Tenerife, Spain
Design and implementation of production backends with FastAPI + LangGraph for public procurement systems of the Madrid City Council, integrating conversational AI agents on Google Vertex AI (Gemini 2.5 Pro) with real-time streaming.

Construction of an automated legal document processing pipeline combining OCR with Docling, metadata extraction with LLMs, and duplicate detection with MinHash + LSH, reducing comparison complexity from O(N²) to O(1).
Implementation of ReAct agent architectures with OAuth2 + PKCE authentication, Fernet encryption at rest, and Passkey login flows integrated with Apigee APIs.
Configuration of CI/CD pipelines in GitLab with automated semantic versioning, builds in Google Cloud Build, and deployments to Cloud Run in multiple environments (dev/staging/prod), with secrets managed via GCP Secret Manager.
End-to-end ML study to automate the routing of citizen requests among 20 administrative units of the Madrid City Council, training and evaluating multiple classifiers on 118,000+ samples with 5-fold cross-validation; 92.8% accuracy and 83.7% F1-macro (+26.5pp over baseline).
FastAPI Google Cloud CI/CD LangGraph Python
Nartex Software
Machine Learning Engineer
August 2023 - September 2025 (2 years and 1 month)
Tenerife, Santa Cruz de Tenerife, Spain
Development of a Retrieval-Augmented Generation (RAG) pipeline with Python, HuggingFace Transformers, and FAISS for regulatory compliance document classification, achieving 93% accuracy.

Orchestration of end-to-end ETL pipelines in Python (Pandas, NumPy) with Apache Airflow to ingest and transform 3 million daily records from REST APIs and PostgreSQL.
Containerization of inference microservices with Docker and deployment on Kubernetes, reducing deployment time by 60%.
Implementation of automatic hyperparameter optimization flows with Optuna, reducing model tuning time by 40% and improving final performance by 5%.
Docker Python Apache Airflow Kubernetes HuggingFace
University of La Laguna
Junior AI Researcher
November 2022 - June 2023 (7 months)
Santa Cruz de Tenerife, Spain
Co-author of a peer-reviewed scientific publication (Computer Methods and Programs in Biomedicine, 2025) on ML surrogate models for clinical decision support systems in type 1 diabetes, achieving IoU > 0.9 with ~2,700x acceleration compared to the original system.
CI/CD Machine Learning PostgreSQL NLP MLOps