A LARGE INVESTMENT BANK MODERNIZED 600+ CRITICAL JOBS WITH AI-LED POST TRADE PROCESS MODERNIZATION

CLIENT

A leading investment bank.

BUSINESS CHALLENGE

The client’s post trade operations relied on 600+ critical batch jobs executed daily after market close. Processing volumes fluctuated dramatically—40M to 100M trades—on a legacy monolithic stack built over 20+ years, comprising COBOL, DB2, and VSAM, and burdened by rising MIPS consumption. They also need to enable fault tolerance support with complex infrastructure for smooth operations.

Due to this aging 15M LOC ecosystem, the bank faced:

High technical debt
Stringent performance expectations (e.g., processing 40M trades within 8 minutes)
Operational fragility requiring fault tolerance and resiliency patterns for smooth MO operations

SOLUTION

Mphasis implemented an AI first modernization program using our proprietary AI platforms:

Mphasis NeoZeta^TM — AI Agent for Legacy Relearning
Used for automated reverse engineering and accelerated understanding of the 15M LOC monolithic codebase.
Mphasis NeoCrux^TM— Modern Engineering Platform
Enabled end to end architecture design, value stream creation, observability, and code development for the new target platform.

Architecture & Engineering Highlights:

Designed a next generation distributed stream processing architecture using Apache Flink

Executed a performance POC demonstrating Flink based high velocity processing

Utilized GPU accelerated Flink execution to meet demanding SLAs

Engineered the full workflow—from file ingest to trade processing to elastic scaling

BENEFITS

The AI driven modern design delivered meaningful and measurable outcomes:

Successfully modernized 600+ critical post trade jobs

Achieved high performance, fault tolerance, and low latency required for mission critical MO operations

Enabled future scalability and operational resilience