Sai Saran

Project: Digital Validation Twin (AI Chatbot Platform) • Designed a production-grade AI chatbot platform on AWS ECS Fargate integrating an Enterprise LLM Gateway (GPT-4.1) to automate project validation workflows and reduce manual review cycles. • Engineered prompt engineering strategies and LLM API integration patterns to enable the chatbot to interpret project specifications, answer domain-specific queries, and deliver accurate real-time CID project status updates via natural language. • Built a high-concurrency backend with FastAPI (Python) and a responsive React.js frontend, enabling the system to handle simultaneous AI-driven requests without performance degradation. • Automated real-time document synchronization from SharePoint using webhook-triggered AWS Lambda functions, ensuring the LLM always operated on the most current project data and eliminating stale-data errors. • Designed and managed PostgreSQL schemas for complex project datasets, ensuring data integrity and enabling fast analytical queries across validation records. • Deployed scalable microservices infrastructure using AWS ECS Fargate, ALB, and Lambda, achieving high availability and fault tolerance for a mission-critical enterprise AI platform.

Client: Guardian Life Insurance Company (Benefits Portal) • Delivered a high-performance insurance benefits portal using React.js, Redux, and GraphQL, enabling employees to manage life, dental, and disability coverage with an intuitive self-service experience. • Integrated SonarQube quality gates and enforced code quality standards, reducing code smell and technical debt by establishing automated quality checks in the CI/CD pipeline. • Containerized frontend and backend services using Docker, and implemented end-to-end CI/CD automation with Jenkins and reducing deployment time and manual release risk. • Collaborated with UX designers via Zeplin and InVision to implement design specifications with high fidelity, reducing design review iterations.

Clients: DFEP (Internal) | Schneider Electric | Pella Corporation • DFEP — Built a micro-frontend platform with React.js and Node.js enabling multiple enterprise apps within a unified experience; implemented JWT authentication, Redux Thunk middleware, and a custom API layer for CRUD operations. • Schneider Electric — Developed a Next.js SSR logistics portal with NATS Streaming for event-driven microservices and Elasticsearch for real-time shipment search; deployed on Azure with Kubernetes orchestration. • Pella Corporation — Built a Gatsby-powered e-commerce storefront integrated with Contentstack (headless CMS) via GraphQL, enabling content-driven product pages for a national window and door brand. • Applied TDD with Jest and Enzyme across all three projects, enforcing test coverage requirements and ensuring production stability through automated regression suites.