Rajat Kumar Dabas

Owned backend architecture for Full Swipe Offer, designing high-throughput data pipelines consuming 20,000 Kafka events/sec and powering real-time merchant analytics via Elasticsearch, MySQL (Aurora), and TimescaleDB; migrated APIs from JSON to Protocol Buffers, achieving 35% payload reduction, 15% lower latency, and 20% throughput increase.

Designed and enhanced a Java/Spring Boot-based distributed rate-limiting system for an accounting microservice, implementing four adaptive sleep strategies to handle 10x traffic spikes while maintaining 99.9% availability. Built and deployed MCP servers and AI-powered chatbots using Python, LangChain, Vercel SDK, integrating OpenAI APIs to automate workflows and reduce external API dependency by 99%. Owned end-to-end backend implementation for summary, listing, and settlement APIs powering the Full Swipe Offer feature, consuming Kafka events and persisting data in Elasticsearch, MySQL, and TimescaleDB with optimized indexing for fast analytics. Migrated 10+ high-traffic APIs from JSON to Protobuf, reducing payload size by 35%, lowering latency by 15%, and improving server throughput by 20%.

Integrated session-based authentication with automated security testing and continuous deployment for 100,000+ user sessions, achieving zero security breaches and 99.9% session persistence for Homeland Group; delivered real-time monitoring dashboard with React Flow visualization and automated CI/CD workflows managing 10,000+ business workflows, reducing workflow errors by 45%.