Avinash Kumar

I’m a Senior Software Engineer and Technical Lead with over 5 years of experience architecting, developing, and deploying highly scalable, fault-tolerant distributed systems. I focus on system design, resilience, and measurable outcomes—reliability, latency, and operational efficiency.

At Revfin, I architected an LLM-powered customer support agent using a RAG pipeline, creating vector embeddings to automate responses for over 140,000 monthly collection and support messages. I also built a no-code voice agent platform for automated collections and tele-verification, handling over 150,000 monthly calls and verifying 2,000+ new loans each month. In parallel, I designed the company’s core IoT data platform that processes over 1 million events per minute and feeds a massive data warehouse (ClickHouse) for instant analytics used in loan decisions and EMI collection strategies.

I spearheaded the migration from monolithic services to a cloud-native architecture on Kubernetes with an Istio service mesh, reducing deployment-related downtime by 60% and improving system resilience. I mentored a team of 8 engineers on microservices architecture, Go, and Node.js, and I established new code review standards. I also automated the software development lifecycle with robust CI/CD pipelines using Bitbucket Pipelines and AWS CodeDeploy, achieving a 95% deployment success rate and reducing manual intervention by 80%.

Earlier, I designed a high-availability payment gateway in Go, processing over ₹50 crores monthly with 99% uptime while ensuring compliance with financial regulations. I improved observability with OpenTelemetry, Prometheus, and Grafana, reducing MTTR by 40%, and I engineered real-time, event-driven processing using Apache Kafka to enable new financial products. I’ve also worked on scalable eSign/eMandate services, re-architecting MongoDB with sharding and compound indexing to reduce critical query latency by 50%, and building backend services with Spring Boot and third-party financial integrations; even as an intern, I developed a Neo4j-based recommendation engine and implemented OAuth 2.0 authentication with Spring Security.