VaultSpeed is an agentic data modeling platform. It is built by a team that has automated data warehouses for enterprises for years, from Leuven in Belgium and from the United States.
Leadership
The people who set the direction
Patrick Van Deven
CEO
Jonas De Keuster
Head of solution engineering
Tom Verschueren
CTO
Mieke Beelen
Head of marketing
Dirk Vermeiren
Co-founder and innovation
Nathalie De Ceulaer
CFO
Koen Moermans
Head of professional services
Where we start
Enterprise data is spread across many systems, and the logic between them is the fragile part
Large enterprises run on many applications. Their data is spread horizontally across ERPs, CRMs, policy platforms, operational databases, SaaS tools, and the systems of every company they have acquired. Each of those systems keeps changing.
The logic that turns that raw data into usable business information sits in between all of them. Over time it becomes the enterprise's most fragile and least governed asset.
The problem we were founded to solve
Transformation debt
Transformation debt is the accumulating cost of transformation logic that is written by hand, undocumented, dependent on a few people and hard to change. Every reader of this page has already built something to cope with it, usually under pressure, and that work deserves recognition. It is also why the debt is hard to see from the inside.
Hard coded logic
Business logic gets hard coded into SQL scripts, stored procedures and ETL jobs instead of being captured in an explicit data model.
Documentation drift
Documentation drifts, or never gets written.
Key person risk
Knowledge concentrates in a few people, and leaves with them.
Unmapped layers
Every new source, schema change or acquisition adds another layer that nobody fully maps.
Why it compounds. Each source multiplies the mappings, each migration reworks logic that was never designed to move, each audit exposes lineage gaps that take weeks to trace. The cost rises exactly when the business needs to move fastest: AI initiatives, M&A integration, cloud migration, regulatory deadlines.
Most data platforms get built twice, once as a data model and again as the code that moves the data. The two drift apart from the first sprint.
The shortest version
Where we come from
Founded in Leuven in 2018, by people who automated Data Vault by hand first
VaultSpeed was founded in 2018 in Leuven, Belgium, by data warehouse automation veterans who had automated Data Vault warehouses for enterprises long before agents existed. Every part of the platform exists because someone on the team once did that work by hand and knew where it broke.
Since 2018 we have maintained one library of templates across all customers. That accumulated knowledge of how enterprise data platforms actually get built is what the platform stands on today.
2018
Founded in Leuven
By data warehouse automation veterans who had done the work by hand for enterprises.
The start
Automating Data Vault
Generating the integration code from the data model. Same input, same code, every run. That work continues today as one of the generators on the platform.
Since day one
One library of templates
Shared by every customer, so each new customer starts with the lessons accumulated across all of them.
Today
An agentic data modeling platform
AI brought to data modeling, with the code still generated the same way every run.
What changed
AI made the transformation layer the weakest link
AI can only be as trustworthy as the data it reasons over, and in most enterprises the transformation layer is the weakest link. Teams now build with AI agents, and agents are becoming cheap and everywhere. What an agent needs before it can act safely on enterprise data, which is data models, lineage, business logic and source knowledge in one governed graph, takes years to build.
That is why VaultSpeed became an agentic data modeling platform: bring AI to data modeling, and automate the code.
Nothing holds between sessions. Every session starts from zero.
Transformation debt, at AI speed.
What we believe
Eight principles we build on
Building on the shoulders of giants
From day one, every customer has run on the same body of code templates and the same rule engine. Dozens of data warehouses are automated with VaultSpeed on that one shared set, so each new customer starts with the complexity and the lessons accumulated across all of them. Because generation is deterministic, that quality holds and can be verified at scale.
The data model is the source of truth
The code is generated from it, so the two stay aligned.
Predictable where the pattern allows, agentic where it does not
Known patterns, Data Vault first among them, get generation driven by rules that writes the same code every run. Where no fixed pattern exists, AI works from pattern libraries, and a human reviews before anything ships. This split is how AI stays usable in regulated environments.
Your data stays in your perimeter
The platform runs in your own cloud, with the LLM of your choice.
You choose the methodology; the platform builds it
3NF, Data Vault 2.0 and others.
Engineers keep their tools
We make them faster their way, from the coding environment they already use.
Your context is yours
Open to any agent, portable across platforms.
A data platform is a living system
Everything is versioned in Git. When sources or rules change, you regenerate instead of rewriting.
Where we are
Leuven, London and Boston
Headquarters in Leuven, a short walk from the historic home of Stella Artois. Offices in London and Boston, with teams on both sides of the Atlantic.
Headquarters
Leuven
Vaartkom 41 3000 Leuven, Belgium
United Kingdom
London
167-169 Great Portland Street, 5th Floor London W1W 5PF, United Kingdom
United States
Boston
200 Portland Street Boston, MA 02114, United States
Join us
Build the data platform that agents and people both rely on
We are a small team whose product runs in production at enterprises on both sides of the Atlantic, and there are plenty of hard problems left to solve. If that sounds like your kind of work, we would like to hear from you.
