Sarvesh
Sawant.

Rebuilt insurance portals in React and Node.js, scaling capacity to ~2K concurrent user sessions through peak monthly premium-payment windows without a single timeout. Decoupled monolithic claims workflows into scalable Java and Spring Boot services, processing ~25K transactions daily with zero downtime. Profiled production execution paths and rewrote poorly indexed PostgreSQL queries, reclaiming ~3 hours of weekly manual verification for underwriting teams. Led the deployment of internal GenAI microservices across 3 engineering teams, containerizing open-source LLMs on AWS Fargate to isolate data. Built a Python pipeline that streams unstructured policy documentation into Amazon OpenSearch vector indexes to power Retrieval-Augmented Generation.

Built internal test-monitoring dashboards in JavaScript and Angular to track telemetry health, device connectivity, and validation status across 40+ bedside monitoring systems. Automated deployment and regression validation with Jenkins CI and Docker within healthcare releases, cutting ~9 engineering hours of manual validation monthly. Analyzed packet-level communication with Wireshark and TCP/IP inspection to validate alarm propagation across 120+ connected telemetry endpoints. Validated wireless roaming and AP failover on IntelliVue Smart-hopping 1.4 GHz WMTS via Azure DevOps across 50+ monitoring devices. Partnered with QA and infrastructure teams using Bash automation and Linux system-testing workflows.

Architected a B2B EDI monitoring dashboard in React and TypeScript that instantly retriggered ~150 daily transaction failures, accelerating recovery. Engineered Kafka-driven event workers that invalidated stale Redis cache entries, eliminating ~50K unnecessary MySQL reads per day. Designed secure persistence layers with Spring Data JPA and OAuth 2.0 token-based authorization, cutting client integration onboarding by ~3 days. Enforced AWS IAM least-privilege across 12 production clusters by automating environment-scoped infrastructure on Kubernetes with Terraform. Shifted operations from reactive tickets to automated alerts across 15 core microservices with Prometheus, Grafana, and Alertmanager.