Harshit Singh

Built production-grade multi-cloud deployment infrastructure reducing customer onboarding time by 96% across AWS, GCP, and air-gapped environments. Delivered 8 components spanning Helm chart libraries, Terraform modules, and hardened Bash automation.

• ›Designed aurva-helm-commons, a centralized Helm chart library orchestrating 15+ microservices across any Kubernetes cluster with IRSA/Workload Identity federation and AWS Secrets Manager CSI integration — reducing new environment deployment from requiring a Helm expert to editing a single values file.
• ›Built end-to-end AWS control plane provisioning (Terraform + Helm) covering VPC, EKS, RDS, OpenSearch, IAM, S3, Secrets Manager, and CloudTrail — including full air-gapped mode via AWS PrivateLink VPC endpoints. Fresh deployment time dropped from 2-3 days to under 45 minutes.
• ›Authored cross-account IAM Terraform module using OIDC-based IRSA federation with aws:SourceVpc conditions, eliminating long-lived credentials and reducing a 2-3 hour manual setup to a 5-minute terraform apply.
• ›Implemented GCP data plane deployment with Workload Identity, GCS bucket provisioning, and optional tainted node pool attachment to existing GKE clusters — enabling multi-cloud support without per-customer manual playbooks.
• ›Wrote hardened Bash installers (eBPF collector, DBProxy, PostgreSQL 18) for standalone Linux servers with preflight validation, retry logic with exponential backoff, rollback, and dual CDN/air-gapped download paths — cutting standalone server onboarding from 4-8 hours to under 10 minutes.

Architected real-time SSH connection pooling achieving 115% performance boost (8K to 17K req/s). Built gRPC gateway layer for hardware communication and led full-stack OpenTelemetry instrumentation cutting debugging time by 40%.

• ›Architected and implemented real-time SSH connection pooling system for remote Linux servers, achieving 115% performance improvement (8K to 17K requests/sec) and reducing resource overhead.
• ›Built scalable high-throughput gateway layer with gRPC streams for seamless hardware communication and efficient API-to-Linux command conversion across distributed infrastructure.
• ›Led comprehensive OpenTelemetry implementation for distributed tracing and metrics collection, reducing system debugging time by 40% and improving operational visibility.
• ›Established standardized protobuf and gRPC best practices across multiple engineering teams, improving code maintainability, development velocity, and reducing integration complexity.
• ›Optimized critical system architecture for high-availability deployment scenarios, ensuring consistent 99.9% uptime across production environments and handling peak traffic loads.

Engineered Redis-based distributed caching cutting database load by 40%. Designed RabbitMQ event-driven architecture improving reliability by 30% and built Docker Swarm orchestration reducing service downtime by 50%.

• ›Engineered comprehensive Redis-based distributed caching system with intelligent cache invalidation strategies, reducing database load by 40% and significantly improving application response times.
• ›Designed and implemented robust event-driven architecture using RabbitMQ message queues, increasing overall system reliability by 30% through decoupled logging, mailing, and notification services.
• ›Orchestrated containerized microservices deployment using Docker and Docker Swarm with automated health checks, achieving 50% reduction in service downtime during critical deployments and updates.
• ›Developed high-performance gRPC communication protocols with auto-generated client-server stubs and connection pooling, improving inter-service communication efficiency by 20% and reducing latency.
• ›Built and deployed automated Go cron scheduling module for incremental billing operations and URL crawling tasks, increasing scheduling accuracy by 25% and reducing manual intervention.