Enrichment Pipeline

Data becomes progressively enriched as it moves from financial institutions into Atlas

As data moves from custodians into Atlas it undergoes several transformations and normalizations. We refer to this as the enrichment pipeline because each stage adds value to the data.

This pipeline begins with Luca, our in-house portfolio accounting system.


Luca is a robust portfolio accounting system that regularly produces less reconciliation and discrepancy than other industry leaders. It scales to millions of accounts and can churn through data for thousands of accounts in minutes.


There is no industry standard for transmitting portfolio data. Transaction codes, reporting methodologies and formatting are just a few aspects of custodian data that differ among institutions.

Luca consumes so-called Lifted data. LIFT (Luca Input Format Translator) is a data preprocessor that eliminates these details by normalizing the data into a common format and reporting practice.


Luca is able to achieve accuracy because it's transactions-based and stores transactions in a ledger-based accounting systems. This means we treat a portfolio the same way an accounting system would treat the receipts for a business. Likewise, just as a corporate accounting system is double-entered to faciliate accuracy of financial statements and auditability of a business, Luca is ledger-based to faciliate accurate and auditable accounting of lot-based financial transactions.

Other systems in the industry tend to be positions based, which means the transactions aren't reliable and subject to large and on-going reconciliation issues.

Reconciliation and QA

It's atypical for positions derived by Luca and those reported by the financial institution to differ. When they do it triggers a reconciliation process. While rare, this situation can arise due to a mismatch in market data or information that hasn't been reported by the custodian (e.g., cost basis).

The reconciliation status is disclosed on each account. It's recommended to check this information regularly to make sure the data you're interacting with has been reconciled.

Sleeve Partitions

Still under development, an account partition corresponds to a sleeve, which is an implementation of an investment model. Housing multiple sleeves within a single account is a way to account and report for multiple investment strategies in one account.