O

Oracle Financial Services

Software Developer • July 2023 - August 2025 • Bangalore, India

Embedded in the Financial Services domain, serving State Street Corporation — one of the world's largest custodian banks managing over $40 trillion in assets. Built enterprise-scale distributed systems processing 10M+ daily transactions.

The core architectural shift: from disk-heavy MapReduce to Spark's in-memory processing model.

Before: Hadoop MapReduce

Read from HDFSDisk I/O
↓
Map Phase
→ Write to DiskI/O bottleneck
↓
Shuffle Phase
→ Write to DiskI/O bottleneck
↓
Reduce Phase
→ Write to DiskI/O bottleneck
Disk I/O between EVERY stage = massive bottleneck

After: Spark on Databricks

Read from S3/Delta Lake
↓
Transform 1In memory
↓
Transform 2Still in memory
↓
Transform NStill in memory
↓
Write final output only
Data stays in RAM → Minimal disk I/O → 20% faster

Additional Technical Contributions

Event-Driven Systems with Scala & Kafka

Real-time streaming, 10M+ daily transactions

CI/CD & Testing Infrastructure

85% coverage, 40% faster deployments

Back to Portfolio