ManiKumar Chintha

About

Backend/full-stack Java engineer at PNC Bank specializing in real-time fraud detection systems. Built event-driven Spring Boot + Kafka microservices with PostgreSQL/Redis performance tuning, and shipped a production LLM-powered RAG feature for fraud analysts with strong guardrails (grounded internal data, structured prompts with references, human-in-the-loop) plus an evaluation loop using labeled historical fraud cases.

Experience

Education

Key Strengths

• Owned end-to-end real-time fraud transaction monitoring feature (build → launch → iterate)
• Low-latency, high-throughput backend design using Spring Boot microservices + Kafka
• Performance optimization via PostgreSQL query tuning and Redis caching to reduce latency
• Strong tradeoff management between response time and fraud detection accuracy
• Iterative post-launch improvement using analyst feedback and metrics (latency, false positives)
• Shipped production LLM/RAG fraud analysis workflow grounded in verified internal data with citations
• Implemented LLM guardrails: prompt constraints, validation checks, and human-in-the-loop review
• Built LLM evaluation loop with labeled historical cases; improved prompts/retrieval to reduce generic outputs and cut analyst review time
• Designed scalable pipeline for messy multi-format transaction data with parsing/normalization and observability (logging/monitoring/alerting)