Owning and leading technical architecture for lender-facing integration services and mission-critical loan lifecycle workflows.
Workflow & Systems Engineering
- Architected flexible gold release and renewal workflows using Camunda Zeebe with Spring Boot and Node.js microservices, boosting customer retention by 30% and revenue by 20%.
- Resolved high-scale concurrency challenges by implementing Redis Clusters and leveraging Lua scripting to ensure atomicity across distributed transactions.
- Engineered and led the end-to-end delivery of the Lender Integration Service, designing a scalable communication layer for external lender ecosystems.
Infrastructure & Security
- Architected a dedicated Pre-Production environment to streamline sanity testing and observability, ensuring 100% safer production deployments.
- Implemented SAML 2.0 SSO integration for Lender Partners, reducing redundant logins and authentication failures by 30%.
- Led zero-downtime migration of legacy JavaScript services to Java Microservices using a Facade-based traffic-routing pattern.
JavaSpring BootNode.jsCamunda ZeebeRedis ClusterLuaSAML 2.0Kafka
In this role, I shifted focus toward architectural ownership and AI-driven optimization, building systems that balanced high-throughput reliability with intelligent user engagement.
Intelligent Systems & AI
- Engineered a search system using pgVector and OpenAI embeddings that utilized query rewriting and taxonomy-based filtering to improve knowledge discovery efficiency by 40%.
- Piloted Small Language Models (Qwen, Gemma) to auto-generate personalized communications, resulting in a 7% increase in click-through rates.
Distributed Systems & Orchestration
- Developed a priority-driven event orchestration system (inspired by Meta's FOQS) to manage large-scale asynchronous workflows, improving SLA adherence by 35%.
- Architected a system using Kafka and Dead Letter Queues (DLQ) that reduced Turnaround Time (TAT) by 70% by moving away from synchronous REST dependencies.
Business Impact (Project 'Rapid LR')
- Led the development of a scalable data streaming architecture using Java, Spring Boot, and Apache Kafka.
- Deployed a modular system that integrated 8 distinct workflows, leading to a 25% growth in MAU, increasing lead conversions by 35%, and achieving an exceptional 98.6% increase in conversion rates for a critical business workflow.
JavaSpring BootKafkaPostgreSQLpgVectorOpenAI APIsAzure Cosmos DbDocker
Focused on platform-level throughput improvements and reusable service tooling for multiple backend teams.
Core Excellence
- Eliminated Azure Cosmos DB connection instability by shifting from Direct TCP to Gateway (HTTPS) mode, successfully resolving partition-level re-acquisition timeouts and stabilizing production throughput.
- Created a reusable TimeSeries DB SDK adopted by 15+ services, reducing integration effort and time-to-market by 20%.
- Increased throughput by 30% through query and indexing optimizations on Azure Cosmos DB.
- Processed 90 million records efficiently using Java multithreading for geo-tagging and data enrichment.
- Elevated overall Collections efficiency by 30% through strategic data initiatives.
JavaCosmos DBAzureMultithreadingMicroservicesKafkaDocker