Enterprise data, reconstructed in time
A lakehouse-native reference architecture for governed data, master data management, distribution intelligence, and enterprise AI — built first for Databricks.
The question most platforms cannot answer
Ask a data platform what a number was on a past date and it will usually tell you what it believes today. That is one answer to two different questions, and the difference is where audits, restatements and disputes live.
What was true on a business date and what the system knew at a point in time are separate facts. A late-arriving correction can be effective from one date while going unknown until a later one. AsOfData treats that separation as the design centre rather than a reporting afterthought.
The bitemporal contract
Golden records, cross-references, value mappings, hierarchies and key stewardship decisions
carry four timestamps: effective_from and effective_to for business
validity, recorded_from and recorded_to for system knowledge.
Standard queries accept both temporal parameters, so both questions have an answer:
as_of
Known — known_as_of
The pair is encoded the same way everywhere it appears — solid border and calendar glyph for effective time, dashed border and history glyph for knowledge time. The distinction is never carried by colour alone.
What this is
The productised asset is the deployable reference architecture, the reusable domain framework, the metadata-driven application framework, and the Distribution distribution-intelligence framework.
The reference implementations are proof, not product. They exist to demonstrate portability and completeness end to end, and they run on synthetic and public-source data. The ETF Security Master is the first of them.
Client deployments use separately isolated environments provisioned from the same Terraform baseline as the reference instance.
Status
Placeholder — release stage, availability and timeline are not stated here yet.
Enquiries
Placeholder — no contact route is published yet.