Biju Lalitha Soman

Modernizing Legacy Financial Platforms: Cloud-Native Replatforming for Extreme Availability and Financial Accuracy

Abstract:

Legacy monolithic platforms in financial services were never designed for today's combination of exponential transaction growth, tightening regulatory expectations, and customer demand for real-time financial visibility. The result is delayed intra-day reporting, throughput ceilings, an expanding blast radius on every release, and reconciliation errors that surface only at end-of-day close—when remediation is most costly.
This talk presents a practitioner's view on how to incrementally replatform these systems incrementally without disrupting live transaction flow. It draws on real-world experience modernizing high-volume payment infrastructure and covers four architectural layers: decomposing the monolith into domain-aligned microservices using domain-driven design and the strangler fig pattern; engineering for five-nines availability through idempotency, circuit breakers, and multi-region active-active deployments; replacing nightly batch cycles with event-driven architectures that make financial state continuously consistent and auditable; and choosing distributed data architectures that satisfy strong-consistency write paths while scaling read-optimized projections independently.
The session closes with a practical playbook for incremental modernization — assessing and instrumenting the monolith, identifying bounded contexts, extracting capabilities via the strangler fig, progressively introducing event streaming, and measuring business outcomes such as financial accuracy and reconciliation error volumes as first-class SLIs alongside infrastructure health. The talk is intended for senior engineering leaders and researchers navigating high-stakes platform transformations in financial and other mission-critical domains.

Profile:

Biju Lalitha Soman is a Director of Software Engineering at JPMorgan Chase, where he works on large-scale payment and clearing infrastructure. His professional focus is on distributed systems, event-driven architectures, and the engineering disciplines required to operate financial platforms at extreme availability. He is a named inventor on a USPTO patent in dynamic database sharding. His prior work includes event-driven processing pipelines and distributed data architecture redesigns at major technology and financial firms.