Re: Model Once, Represent Everywhere: UDA (Unified Data Architecture) at Netflix

Hi Alexandre,

Very nicely written post!

It sounds like part of the tooling developed in the process includes tooling to specifically transpile SHACL into schemas and code. I expect many people on this mailing list have or use tools which do this - and would be interested in your solution. Is any of the tooling that you've developed for this publicly available?

Kind Regards,

Jesse Wright
DPhil Computer Science
Jesus College
[cid:b2a0fa2a-1e01-40c1-972b-a3261438623a]

mailto:jesse.wright@cs.ox.ac.uk
https://www.cs.ox.ac.uk/people/jesse.wright/
https://www.linkedin.com/in/jesse-wright-49823a132/
Book a meeting: https://cal.com/jesse-wright-zdbdal/public
Mobile: +44 7862 381 515
WhatsApp: +61 468 669 019

My work day may look different than yours. Please do not feel obligated to respond outside of your normal working hours.
________________________________
From: Alexandre Bertails <abertails@netflix.com>
Sent: Thursday, June 12, 2025 7:46 PM
To: public-shacl@w3.org <public-shacl@w3.org>
Subject: Model Once, Represent Everywhere: UDA (Unified Data Architecture) at Netflix

Folks,

I'm excited to finally share more about Netflix’s motivation for joining this group. Today, we’ve introduced UDA (Unified Data Architecture) [1] to the public:

[1] https://netflixtechblog.com/uda-unified-data-architecture-6a6aee261d8d

> UDA (Unified Data Architecture) is the foundation for connected data in Content Engineering at Netflix. It enables teams to model domains once and represent them consistently across systems — powering automation, discoverability, and semantic interoperability.

This post introduces the foundations of UDA as a knowledge graph, connecting domain models to data containers through mappings, and is grounded in a metamodel we call Upper (short for "upper ontology") which may be of particular interest to this group.

Upper is a self-referential, self-describing, and self-validating modeling language built on top of SHACL. It extends SHACL’s expressivity for domain modeling and introduces additional constraints that support strong typing, projections, and introspection. In our architecture, every upper:DomainModel is, by design, also a SHACL ontology.

We see Upper as a SHACL profile, and equally as a profile of RDF. Any conceptual RDF can be described, as long as it is governed by a domain model defined in Upper.

Beyond modeling, UDA also productionalizes RDF:

* Every piece of data is stored in a Named Graphs, each conforming to a governing model and owned by a team.
* Upper itself is part of the bootstrapping set of domain models, alongside formalizations of RDF, RDFS, OWL, SHACL, and XSD.
* Our internal partners have contributed axiomatizations of SKOS and DCAT using Upper, introduced in the post as well.
* We also introduce the notion of composable universes, where Named Graphs resolve into Governed Graphs via follow-your-nose import predicates declared in governing models.

We’ll be sharing more technical details on Upper in the coming weeks. In the meantime, I’d be happy to answer any questions.

Best,
Alexandre Bertails