• Contributed to the development of Capital One’s Prescreen Decisioning Platform, a large-scale credit pre-approval system processing 1M+ applications daily, helping automate eligibility and risk decisions and improving approval turnaround time by 35%.

• Built Python backend services using FastAPI, implementing APIs for eligibility evaluation, credit scoring orchestration, and decision storage; deployed services on AWS Lambda with API Gateway to support stateless, low-latency request handling.

• Refactored decisioning workflows into containerized services using Docker and AWS Fargate, separating eligibility, scoring, and audit components to improve fault isolation and enable independent scaling during peak prescreen campaigns.

• Developed Apache Spark (PySpark) jobs on Amazon EMR to process and standardize large credit bureau and transactional datasets in Amazon S3, ensuring consistent data inputs for downstream decisioning and analytics.

• Integrated machine learning models hosted on AWS SageMaker into Python services for real-time credit risk scoring, contributing to an 18% improvement in model-driven decision accuracy.

• Built an internal GenAI-assisted tool using Python and LangChain to generate plain-language credit decision explanations from model outputs and rule evaluations, reducing manual review and documentation effort.

• Improved system performance by optimizing PostgreSQL queries and introducing Redis caching for frequently accessed eligibility data, reducing database load and lowering average API response times by ~30%.

• Supported reliable deployments and production stability using Jenkins for CI, terraform for AWS infrastructure, and Datadog monitoring, helping reduce incident resolution time by 40% while maintaining PCI DSS and SOC 2 compliance

• Led the architecture and implementation of Amazon's 3rd-Party Seller Onboarding workflows, modernizing legacy flows into scalable, event-driven pipelines using AWS Step Functions, Lambda, API Gateway, and DynamoDB — reducing onboarding time from 3 weeks

’3 days.

• Migrated high-volume financial reporting pipelines (fee transactions & settlements) to modern microservices, improving data consis- tency, auditability, and operational reliability across Amazon's internal finance tools.

• Improved API response latency by 25% by redesigning internal workflows, removing bottlenecks, and introducing intelligent caching layers.

• Optimized the Amazon Pay Dashboard pipeline, using CloudWatch metrics, CPU/memory profiling, and infra-level debugging to improve throughput and system resilience during peak load.

• Designed and executed microservice migrations from a legacy monolith, improving horizontal scalability, deployment isolation, and failure containment for high-traffic seller systems.

• Built cost-efficient infrastructure blueprints and optimized compute/storage patterns, contributing to 15% AWS cost reduction across targeted services.

• Automated UI workflow validations using Selenium + TestNG, integrating them into CI/CD pipelines for deployment-ready UI checks and zero-regression releases.

• Enabled UI theming on the Amazon Shopping Gateway, contributing to global content campaigns and improving customer-facing content customization.

• Played a key role in 24/7 on-call rotations, performing root cause analysis, resolving customer-impacting incidents, reducing MTTR, and improving system observability.

• Architected and delivered a multi-tenant Learning Management System (LMS) as part of a larger microservice ecosystem, capable of handling 800+ RPS. Designed for high concurrency, memory-safe request handling, and autoscaling using Django, Node.js, PostgreSQL, AWS EC2, and distributed caching - reducing response latency by ~30% and enabling rapid onboarding of new enterprise clients.

• Built the platform's OAuth 2.0-based authentication system with JWT + refresh tokens, improving login reliability and strengthening platform security by 3x.

• Led a team of 6 engineers, owning sprint planning, code reviews, architectural decisions, and technical mentorship for interns and junior developers, improving overall delivery velocity.

• Implemented automated end-to-end test suites using PyTest, TestNG, and LambdaTest for cross-browser and mobile compatibility, integrating them into CI/CD pipelines with GitHub Actions and Jenkins.

• Developed core backend modules including user onboarding, course management, analytics dashboards, and RBAC permission systems, contributing to strong customer adoption.

• Optimized database models, indexing strategies, and asynchronous workers, improving backend throughput by ~40% and ensuring platform stability during traffic spikes.