Acerca de Javier
I help companies build production-ready GenAI systems — not proofs of concept, not demos, but pipelines that run at scale and deliver real business value.
As GenAI Tech Lead at Zurich Insurance, I architect LLM-powered systems that process 10,000+ documents monthly, extracting hundreds of structured variables per document. I lead teams of up to 10 engineers and own the full lifecycle: from architecture design to deployment and monitoring.
What sets me apart: I've shipped agentic AI pipelines in production using LangChain, LangGraph, LangFuse, and RAG — not just experimented with them. My background spans Amazon (predictive analytics across 50+ EU fulfillment centers), a CTO role at an AI startup, and 7+ years bridging ML research with production engineering.
What I deliver:
→ LLM-powered extraction and classification pipelines
→ Agentic AI systems (multi-agent orchestration with LangGraph)
→ RAG architectures with vector databases
→ Cloud infrastructure design (AWS, GCP)
→ MLOps/LLMOps: CI/CD, monitoring with LangFuse, Terraform IaC
→ Technical leadership and team mentoring
Stack: Python, LangChain, LangGraph, LangFuse, RAG, PyTorch, FastAPI, Pydantic, Docker, Terraform, AWS, GCP, MongoDB, SQL.
Previously: Amazon, Zurich Insurance, CTO at Nitcai. Aerospace Engineer (UPM) + Master in Big Data.
Español
Bilingüe o nativo
Inglés
Competencia profesional completa
Acepta trabajo presencial
Marbella (hasta 50 km)
Experiencia
- Zurich InsuranceGenAI Tech LeadHIGH TECHseptiembre de 2024 - Hoy (1 año y 11 meses)SpainLead a team of up to 10 engineers across two document understanding projects using LLMs. Own technical strategy, architecture, and delivery for business-critical data extraction systems.Policy Document Intelligence– Designed and deployed LLM extraction pipeline processing 5,000–10,000 policy documents/month, extracting 200–500 structured variables per document– Replaced legacy manual systems for detecting discrepancies between master and local policies, eliminating coverage misalignment that previously required full manual review– Transformed unstructured policy data into structured KPI feeds for business analytics and compliance monitoring Clause Detection & Language Quality– Built system to identify specific legal clauses and extract relevant text snippets from policy documents– Developed quality scoring framework assessing language clarity and regulatory complianceTechnical leadership product development
- LiveWell by Zurich — Zurich InsuranceTech LeadHIGH TECHmayo de 2022 - enero de 2024 (1 año y 8 meses)SpainPromoted from Senior DS to Tech Lead. Led teams of 3–5 building AI-powered wellbeing features for a consumer health platform serving thousands of monthly active users.– Designed and shipped Generative AI content feature delivering personalized wellbeing recommendations– Built NLP validation layer (TF-IDF, embeddings) ensuring LLM output quality before user-facing delivery– Built end-to-end recommendation engine using BERT embeddings for personalized content delivery– Designed DynamoDB schema and serverless API (AppSync + Lambda) powering the recommendation layer– Architected data model on DynamoDB + Redshift Serverless; defined all infrastructure as Terraform IaC with CI/CD on GitHub Actions
- AmazonData Scientist II — EU RME Predictive Analyticsjunio de 2020 - agosto de 2021 (1 año y 2 meses)RemoteBuilt predictive maintenance and NLP systems for Amazon's Reliability & Maintenance Engineering team, analyzing data across 50–60 European fulfillment centers.– Designed predictive maintenance models (XGBoost) for industrial machinery to reduce unplanned downtime across EU operations– Built NLP system to detect temporal patterns in maintenance work orders, improving scheduling efficiency– Conducted exploratory analysis on terabyte-scale datasets spanning the majority of Amazon's EU fulfillment network– Developed NLP tools for automated data quality audits of maintenance work orders
Recomendaciones
Sé el primero en recomendar a Javier
Ayuda a este freelance a destacar compartiendo tu experiencia.
Estos perfiles de freelance también coinciden con tus criterios
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Formación
- Master — Big Data and AnalyticsMBIT School2019Master — Big Data and Analytics
- Bachelor of Aerospace EngineeringUniversidad Politécnica de Madrid (UPM)2017Bachelor of Aerospace Engineering