Sai Kiran Challa

Refactored core train-tracking platform into 15 independent microservices using Kafka as the sole integration backbone, eliminating direct service-to-service HTTP calls and maintaining sub-200 ms latency during signaling disruptions. Architected high-performance REST and gRPC APIs for passenger apps and control-room systems serving 1.8M+ daily users. Implemented Redis Streams-based real-time disruption alerts, reducing passenger notification delay from 12s to under 800ms. Reduced PostgreSQL load by 65% via read replicas, materialized views, and prepared statement caching. Deployed Resilience4j circuit breakers and Istio policies to isolate faulty feeds in safety-critical environments. Enabled zero-downtime deployments with Argo CD canary releases and Prometheus-driven rollbacks (3+ releases/week, 100% success). Implemented end-to-end distributed tracing with OpenTelemetry and Jaeger across microservices, cutting MTTD/MTTD by ~55% and speeding recovery during peak faults.

Designed and deployed high-transaction Java microservices processing 2M+ API requests/day with 99.9% uptime and latency under 200ms. Optimized REST APIs with Spring Boot for real-time gaming transactions across U.S. states, reducing downtime by 40%. Engineered event-driven messaging using Kafka and RabbitMQ handling 5M+ daily events with fault-tolerant delivery. Improved throughput by 35% through multithreading and connection pooling; resolved messaging bottlenecks via profiling, increasing reliability by 60%. Built fault-tolerant pipelines with retry and dead-letter queues achieving zero message loss and 75% higher reliability. Integrated MongoDB and PostgreSQL with caching to cut complex query times by 55%.

Delivered Java 17 / Spring Boot microservices handling 2M+ financial transactions/day with <120 ms p95 latency and 99.999% availability across Australian jurisdictions. Implemented scaled Saga orchestration with compensating transactions, removing two-phase commits while ensuring regulatory compliance. Consolidated DynamoDB into a single-table design with sparse indexes, reducing monthly AWS costs by >70%. Optimized virtual threads and lock-free collections, boosting peak throughput from 8k to 13k RPS. Designed Envoy + Lua-based adaptive rate-limiting gateway protecting REST/gRPC endpoints during 10x traffic spikes. Rolled out Pact consumer-driven contract testing across 30+ microservices, reducing production integration defects by ~90%. Led Prometheus + Grafana dashboards for 50+ microservices and batch pipelines. Implemented idempotent payment workflows using the outbox pattern, eliminating double charges and reducing reconciliation mismatches by >90%.