Scaling Reinsurance Compliance: The Engineering Behind PartnerRe’s "Overlay"
Details
To meet regulatory standards like Solvency II and LDTI, reinsurance firms must simulate how thousands of insurance policies perform under thousands of potential economic futures.
In data terms, this creates a "Cartesian explosion"—multiplying two manageable datasets creates a result set of billions of rows. This massive computational workload frequently overwhelms traditional on-premise infrastructure.
In this session, Zsolt Revay and Thomas Mager detail the engineering behind "Overlay," the internal data application PartnerRe built to solve this scaling challenge.
Overlay’s architecture decouples business logic from compute power. We use Datavirtuality as a semantic orchestration layer, while leveraging Snowflake strictly as a heavy calculation engine.
By using an aggressive SQL push-down strategy, we treat the database as a compute cluster. This allows us to process that Cartesian explosion of policy data and stochastic scenarios right where the data lives—without moving it across the network.
Key Technical Discussion Points:
- Optimisation: Handling massive joins via parallel processing and strategic table materialisation.
- Performance: Using Elastic Virtual Warehouses to achieve linear scaling, reducing processing time from hours to minutes.
- Economics: Using Auto-Suspend and Resource Monitors to keep this massive compute power cost-effective.
Attendees will learn how to architect a solution that balances the auditability required by regulators with the on-demand agility required by actuaries.
Event Details
📍 14th January | Clayton Hotel
18:00 | Food & Drinks 🍽️🥂
18:30 | Scaling Reinsurance Compliance: The Engineering Behind PartnerRe’s "Overlay"
19:15 | Drinks & Networking 🍻
Don’t miss this opportunity to connect with fellow data professionals and gain valuable insights into efficient Snowflake use, Gen AI applications, and proving the value of solid data modeling! See you there!
Speakers:
Zsolt Revay - Head of Data Engineering
Thomas Mager - Head Data and Analytics Platforms