01
Data Platforms
End-to-end platforms on Microsoft Fabric and dbt: medallion and Data Vault modelling, incremental loading, CI/CD across environments, and monitoring that alerts before the business notices.
Data Analytics EngineerMadrid, Spain
Data platforms people and AI agents can trust.
I turn business knowledge into simple, reliable data models, and bring AI agents into the data platform in a secure, reviewable way.
Currently atSparkle NVJoining Ebury · Nov 2026Toolbox
Tools I build with.
From Microsoft Fabric to my next stack at Ebury: BigQuery, Airflow, Google Cloud, Sigma.
- Microsoft Fabric
- dbt
- BigQueryNext
- Power BI
- Claude Code
- AirflowNext
- Azure
- Python
- Google CloudNext
- Azure DevOps
- PostgreSQL
- SigmaNext
- SAP
- Git
- OpenAI
- LangChain
01Expertise
What I do.
02
Business-Ready Models
Semantic models, Power BI reports and data-quality worklists built with analysts and business users, so the numbers match how the business actually works.
03
AI in the Data Platform
Ontologies and data agents that answer business questions within security rules, and AI coding agents that speed up delivery behind reviews, tests and human-approved releases.
02Experience
Where I've worked.
Starting November 2026
Upcoming
Analytics Engineer
Ebury- Joining a fast-growing fintech for cross-border payments, FX and hedging, to own the data models of a core product domain.
- Bringing my Fabric and dbt experience to a Google Cloud stack: dbt, BigQuery and Airflow, with Sigma for BI.
- dbt
- BigQuery
- Airflow
- Sigma
- Fintech
2024 — Present
Current role
Data Analytics Engineer
Sparkle NV- International trading group: built a finance data platform on Fabric and dbt that unifies several ERPs across countries, using Data Vault, reconciliation against the legacy system, row-level security and CI/CD across four environments.
- Public-sector IT services company: split a monolithic Fabric setup into integration, transformation and consumption workspaces, archived ~275 legacy dbt models and set up central monitoring and alerting.
- Traced half a day per week of manual invoice corrections back to the linking logic, and raised automatic invoice-to-order matching to ~81%.
- Built an ontology and knowledge graph on governed gold models, so a Fabric data agent answers multi-hop business questions within row-level security.
- Brought AI coding agents into daily delivery behind isolated branches, a convention-checking reviewer agent, CI checks and human-approved production releases.
- Microsoft Fabric
- dbt
- Data Vault
- Power BI
- Fabric IQ
- Claude Code
2023 — 2024
Innovation Internship
werecircle- Prototyped chatbots and generative-AI tools with OpenAI, LangChain and LlamaIndex, and competed in the Microsoft OpenAI Hackathon.
- Explored Azure ML Studio (Prompt Flow), AI Studio and web scraping, working independently in a start-up environment.
- Web Scraping
- Gen AI
- Microsoft Azure
- Start-up
03Education
What I've studied.
Postgraduate Programme in Innovation & Entrepreneurship in Engineering
KU Leuven
Specialized program focusing on innovation management, business development, and entrepreneurial skills in engineering contexts.
2023 — 2024
Master's in Industrial Sciences: Internet Computing
KU Leuven Campus Groep T
Advanced studies in cloud computing, distributed systems, and web technologies, with a master's thesis focused on a machine learning topic.
2022 — 2023
Bachelor's in Industrial Sciences: Electronics-ICT
KU Leuven Campus Groep T
Foundation in electronics and information technology, covering programming, networking, and system design. Gained practical experience through hands-on (team) projects and laboratory work.
2019 — 2023
Science & Mathematics
Scholen Kindsheid Jesu
Proper foundation in mathematics, sciences, and analytical thinking.
2013 — 2019
04Approach
How I work.
Understand what is built
AI agents now write a growing share of data code. My job is to know exactly what was built and why, whoever wrote it, and to spend my time on the business problem rather than on boilerplate.
dbt Labs, State of Analytics Engineering 2026: 72% of teams prioritise AI-assisted coding, only 24% AI-assisted testing.Keep models simple
I sit with analysts and business users and turn their daily knowledge into data models. Clear names, one definition per concept and good documentation make a model readable for people and for AI agents alike.
Spider 2.0 (ICLR 2025): LLM agents solve 21% of tasks on complex enterprise schemas vs 91% on simple ones.AI, with guardrails
I went from building everything by hand to working with agents every day. That only works with isolated branches, automated reviews, tests in dev first and a human approval before anything reaches production.
KPMG & University of Melbourne, 2025: 66% of people use AI regularly, but only 46% are willing to trust it.Remote by default, together when it matters
Working from coworking spaces in Madrid taught me that remote work delivers, flexibly and reliably. Trust is still built in person, so I make time to meet the people I build for.
Bloom et al., Nature 2024: hybrid work cut resignations by a third with no loss in performance.Languages
- DutchNative
- EnglishFluent
- FrenchIntermediate
- SpanishIntermediate
Off the clock
- Running
- Tennis & Padel
- Drums
- Percussion