Ranjit Subedi

Designed and implemented credit risk models using Random Forest and XGBoost, reducing delinquency rates by 14%. Integrated AWS SageMaker pipelines for scalable model training and deployment. Analyzed billions of transactions with PySpark and Hive to detect fraud patterns. Collaborated on migrating legacy models to Snowflake, improving processing by 40%. Built Tableau dashboards delivering real-time KPIs on consumer behavior and risk trends. Developed NLP sentiment analysis tools improving customer service responses. Automated model monitoring reducing false positives by 22%.

Architected and developed enterprise-grade full stack applications using ASP.NET Core, React.js, Node.js and Next.js. Built scalable Angular applications using reusable components, RxJS, and NgRx for state management. Implemented AI-powered platforms using LangChain and LangGraph for multi-step reasoning workflows. Developed multi-agent AI systems capable of planning, reasoning, and executing enterprise workflows. Implemented RAG pipelines with hybrid search (PGVector + OpenSearch) for intelligent document retrieval. Built tool-calling frameworks enabling AI agents to integrate with APIs and backend systems. Built scalable backend services using ASP.NET Core and FastAPI. Improved database performance by 30% through query optimization and indexing. Implemented JWT and OAuth2-based authentication with secure API design.

Engineered a high-scale financial dashboard using React 18, TypeScript, Redux, and Redux-Saga for real-time transaction monitoring. Implemented responsive layouts with Tailwind CSS, SASS, and Styled Components supporting theming and branding for multiple clients. Optimized app performance by utilizing Next.js SSR/SSG features, code-splitting, and lazy loading, reducing load times by 25%. Integrated REST APIs for card insights, fraud alerts, and user data with robust error handling and fallback states. Ensured strict WCAG 2.1 accessibility compliance including ARIA roles and screen reader support. Led the migration from Angular to React, improving efficiency and reducing technical debt. Established testing frameworks using Jest and Cypress, reducing manual QA by 40%. Mentored frontend engineers and contributed to agile planning and architecture reviews. Developed an AI-powered Compliance Chatbot leveraging LangChain, LangGraph, and PGVector for real-time internal document search. Design