Open Source in Finance Forum London 2025

Risk and Finance functions are poorly served by modern data science patterns because most of the analytics these functions run are structural and subject to complex control frameworks, including those related to model governance, financial control, data quality and data lineage. Examples of structural analytics include the models which inform regulatory calculations, accounting valuations, pricing and risk-based decisioning.
While modern data science tooling excels at delivering scalability, flexibility and the ability to automate the train model > deploy model process, structural analytics requires a different set of capabilities - control, lineage, repeatability and auditability.
TRAC D.A.P. – a FINOS project – offers a new approach to structural analytics. Built around a universal metadata model, the platform is self-describing and stateful, enabling users to:
-  Upload models and build complex deployments in minutes
-  Fully automate governance documentation
-  Self-serve with confidence, absent any change risk
-  Use the same resources for production and experimental models runs
-  Replicate any prior calculation with ease
This session will explore:
1.     The challenges associated with building platforms for regulatory models and structural analytics
2.     TRAC D.A.P. and the benefits of a universal, stateful and self-describing system
3. How to can leverage and extend TRAC D.A.P. project