Avi Bhardwaj

I work across the whole cloud landscape rather than any single corner of it — AWS, GCP and Azure, with Terraform and Kubernetes as the backbone. Six-plus years where the difficulty lives: reliability, cost, security, and how they hold together at scale. The work is concrete: production serving 1M+ daily users at 99.9% uptime; three Kubernetes clusters consolidated into one multi-tenant cluster with automated DC–DR failover and cross-region migration over WireGuard; an AI startup's spend cut 70%+; SOC-readiness from zero to production, static keys replaced with short-lived OIDC identity, and an open-source MDR stack for centralized detection and response. More recently, AI-native infrastructure: single-purpose agents run live against a real AWS Organization under plan-then-approve control — every action reviewable, every audit checked against external benchmarks, not the agent's own. Automation earns autonomy only when it stays reviewable, and that principle runs through all of it.