Architected a production-grade RAG-based AI support system using LangChain, OpenAI APIs, and FAISS/Pinecone vector databases, enabling 10K–15K queries/day and sustaining 120–180 QPS peak load across distributed support channels with high availability and low-latency performance. Designed scalable microservices architecture using Java and Spring Boot AI services with REST and GraphQL APIs, improving system throughput by 35% under high concurrent load scenarios. Developed LLM-based agentic dispute classification system, integrating PostgreSQL/MySQL pipelines and prompt-engineered workflows, processing 50K+ financial dispute records/month across multiple fraud categories and decision flows. Implemented real-time event-driven processing system using Apache Kafka and Redis caching, enabling streaming of customer queries with optimized retrieval latency reduced to sub-300ms response time. Deployed containerized AI services using Docker and Kubernetes on AWS (EKS, Lambda, S3, API Gateway) with Terraform-based Infrastructure-asCode, improving deployment scalability and reducing infrastructure provisioning time by 50%. Integrated CI/CD automation pipelines using Jenkins and GitHub Actions, incorporating unit testing frameworks (JUnit, Jest) and automated rollback strategies, achieving 85%+ test coverage and 30% reduction in production defects. Engineered a real-time financial anomaly detection engine using graph analytics and streaming pipelines to identify cross-account fraud patterns, strengthening systemic risk visibility across large-scale payment networks under strict fintech compliance. Operationalized observability and monitoring dashboards using Splunk and AWS CloudWatch, enabling real-time tracking of system health, API failures, and AI inference behavior, improving incident detection from delayed logs to near real-time alerting workflows.

Built and enhanced a highly scalable, enterprise-grade digital banking application using Java, Spring Boot, Hibernate, and REST APIs, supporting secure core banking features like fund transfers, account lifecycle management, and loan processing with 99.9% system uptime in production environments. Configured secure authentication and authorization framework using OAuth2, JWT, and AES-256 encryption with TLS 1.2 security protocols, significantly improving API security posture and reducing unauthorized access risks across distributed banking services. Constructed highly responsive, component-based frontend architecture using React.js with Redux state management and integrated backend services built with Node.js and Spring Boot, improving application responsiveness and reducing page load latency by 40% Architected a legacy-to-modern integration layer using API gateway orchestration and service mesh architecture, enabling secure interoperability and seamless data sync across banking systems and third-party platforms in distributed environments. Structured and optimized a real-time, high-throughput transaction processing system using event-driven microservices architecture, Redis caching strategy, and MongoDB NoSQL database, enabling seamless handling of 10,000+ concurrent transactions per minute with sub-200ms latency. Formulated and maintained fully automated CI/CD DevOps pipelines using Jenkins and GitHub Actions, enabling continuous deployment of containerized microservices on AWS cloud infrastructure (EC2, S3, Lambda, API Gateway) and significantly accelerating release cycles. Created comprehensive automated testing frameworks and quality assurance suites using Jest for frontend and Mockito for backend services, achieving 85%+ code coverage and reducing production defects through continuous integration validation and regression testing.