About
A small practice that builds the thing itself.
QualityLake is the data-platform practice of TrendVectors SL. It is founded and led by a senior data engineer and Databricks architect with more than twenty years in the field, and it delivers with a small network of trusted specialist contractors rather than a bench that needs filling.
That is a deliberate constraint. It means we take fewer engagements, the person who designed your platform is the person who writes the code, and nobody is assigned to your project because they were available.
Where we are
Based in València, Spain. We work across Spain and Germany and remotely across Europe, on-site where the work needs a room and a whiteboard.
Languages
Working languages are English and German, both at C1. Design reviews, documentation and workshops run in either.
Education
Computational and applied mathematics, with a focus on probability and statistics. It shows up mostly in how we reason about data quality and sampling.
Certifications
Current, and kept current.
Five active certifications. No logos, no partner badges — just the list.
- Databricks Certified Data Engineer Professional
- Databricks Certified Azure Databricks Platform Architect
- Microsoft Certified Azure Solutions Architect Expert
- Microsoft Certified Azure Data Engineer Associate
- Microsoft Certified Fabric Analytics Engineer Associate
How we work
Five things we will not argue about.
- 01
Data quality is a pipeline property, not a dashboard feature.
- 02
Governance you cannot express in code is governance you cannot audit.
- 03
A platform nobody can onboard onto is a bottleneck with good intentions.
- 04
The cost of a data platform is an architecture decision.
- 05
Fixed scope, written deliverables, no surprises.
We build our own product on the same foundations.
TrendVectors — the news-signal product of the same company, TrendVectors SL — runs on the architecture QualityLake builds for clients: a governed lakehouse, streaming and batch ingestion side by side, quality rules at every boundary, and infrastructure and deployment expressed entirely in code. We operate it ourselves, we pay its compute bill ourselves, and we are the first people to notice when a pattern does not hold. That is the reason we are willing to say these patterns work.