Site Reliability Engineer based in Bangalore.
I keep fintech systems alive — Kubernetes, AWS EKS, Terraform, Oracle SQL.
Turned a 42% API failure rate into 90% success. That's the job.
Clari5 / Bigseer (Vendor Contract) · Bangalore, India
Sep 2023 – Present
Designed, containerized, and deployed a Rocketlane-style internal ticketing tool on AWS EKS using Kubernetes and Terraform — end-to-end infra-as-code ownership.
Managed production monitoring stack with Prometheus and Grafana — dashboards, alerting, and proactive health checks across all services.
Diagnosed and resolved NCRP API integration defect (wrong message type + DA lien amount mapping) — improved API success rate from 42% → 90%.
Oracle SQL query optimization and PL/SQL-level debugging to support fraud alert tracking and reduce data retrieval latency.
Built Python automation for log parsing, API health checks, and scheduled reporting — reduced manual ops effort by ~40%.
Handled L2 incident escalations: RCA, runbook documentation, and post-incident reviews for fraud/AML rule failures.
Developed and validated AML/fraud detection scenarios and rule configurations for banking client environments.
What I've Built
Projects
01
Internal Ticketing Tool — EKS Deployment
Full-stack internal project management tool (Rocketlane-like). Containerized with Docker, deployed on AWS EKS. Infrastructure provisioned via Terraform — zero-downtime rollouts via Kubernetes ingress.
Docker · AWS EKS · Terraform · Kubernetes
02
NCRP API Integration — Defect Resolution
Identified dual root causes in production: wrong SFMS message type + DA lien amount mapping. Fix coordinated across app layer and Oracle DB. Validated via Postman regression. Formal RCA delivered to client.
42% → 90% success rate improvement
03
Sentinel — Fraud Detection Platform
Designed a comprehensive real-time fraud detection platform architecture — Kafka, Flink, ScyllaDB, hybrid ML pipeline. Built React SPA frontend with full dashboard for alert management.
React · Kafka · Flink · ScyllaDB · ML
04
Flight Delay Prediction System
ML model with 83% prediction accuracy using historical and real-time data from weather, traffic, and flight schedule APIs. End-to-end pipeline from data ingestion to API-served predictions.