Shreyansh Jain

Built document crawling and ingestion systems supporting multiple file types and web formats using LangChain with modules like unstructured and PDF parsers — processing 10K+ documents weekly across structured and unstructured sources. • Designed and deployed serverless architecture by migrating workloads from EC2 to AWS Lambda, improving scalability and reducing infrastructure costs by 70% while achieving 3× faster de- ployment cycles. • Integrated AWS Lambda, SQS, and CloudWatch for asynchronous task handling, retries, and real- time monitoring — increasing system reliability and reducing failure rates by 40%. • Created a LangGraph-based MCP server for Aws S3, search documentation, integrated with Zoho ManageEngine to automate ticket creation and lifecycle management — reducing manual support effort by 60% and improving SLA adherence by 35%. • Built a scheduler system to automatically crawl and process multiple data sources (web URLs, sitemaps, and files) — enabling continuous ingestion of 1K+ sources daily with near real-time updates. • Enhancedagentaccuracyandrobustnessbyrefiningpromptengineeringandchaininglogic(LangChain) — improving dynamic query handling accuracy by 25% and reducing model fallback rates by 30%. • Containerized AWS Lambda functions using multi-stage Docker builds and deployed them via Amazon ECR. • Built and maintained three independent Dockerized Lambda services for modular, scalable serverless architecture.