Harsh Rajput

– Legs Platform

• Revamped Causix microservice in Go to detect call types using AI inference, concurrently processing 10K+ calls per minute.
• Orchestrated 11 microservices into a Compose stack, cutting onboarding time from 2 weeks to 1 day for QA and local setups.
• Automated data recovery from PostgreSQL to ElasticSearch using goroutines, reducing effort from 5 days to 30 minutes.
• Implemented IAM roles across 137 servers to replace AWS credentials after a security breach, enabling automatic key rotation.

• Built and scaled a fraud detection microservice in Go, running AI inference to process 1,500+ biometric entries per minute and automate 85% of Aadhaar de-duplication via Kafka pipelines.
• Optimized system performance using MySQL composite indexing and Redis-backed pipelines, achieving sub-50ms cache-miss latency and 30% higher throughput under stress.
• Architected a secure HSM-encrypted biometric data pipeline (S3/Redis) with 80% test coverage, ensuring reliable and low-latency access across services.
• Developed and deployed an operations portal from scratch, reducing manual workload by 75%, cutting DLQ processing time from 30 to 5 days, and improving efficiency by 40% with tracking and prioritization features.
• Designed and deployed an Apache APISIX API Gateway on Kubernetes, managing 9,000+ daily requests, implementing routing, and enabling secure Keycloak-based token authentication in a three-legged auth flow.