Java Backend Engineer specializing in production-grade system design, multi-tenant architecture, and REST API development using Java 21 and Spring Boot 3. Built real-world systems including a multi-tenant SaaS platform for Indian district courts featuring JWT security, FSM-driven workflows, and role-based access control. Seeking a backend engineering internship to contribute to scalable, well-architected systems.
Engineered Lexora, a multi-tenant SaaS for Indian district courts using Java 21, Spring Boot 3, PostgreSQL, shared schema architecture, ThreadLocal tenant isolation, and zero cross-tenant data leakage. Implemented stateless JWT authentication, custom Spring Security filter chain. Designed an FSM-driven case lifecycle with immutable audit logging via Spring domain events, Quartz-scheduled idempotent SMS reminders. Built paginated REST APIs, DTO response filtering, global exception handling, and a manually implemented token bucket rate limiter.
Engineered a robust Java/Spring Boot backend for an intelligent medical assistant, integrating local Ollama LLMs to summarize complex patient records 40% faster. Architected a multi-model Agentic AI pipeline to translate healthcare data, processing 10,000+ simulated clinical entries with 95% summarization accuracy.
Designed and built a production-grade middleware layer that intelligently routes prompts across multiple local LLMs based on complexity, serves semantically similar requests from a vector cache instead of re-running inference, and continuously evaluates model quality using an automated judge-model pipeline. Owned the full system design — gateway, router, cache, fallback chain, and eval pipeline — modeled on the infrastructure patterns real AI infra teams (not chatbot/RAG projects) use in production.