ABOUT PRATHAMESH
I find the edge cases in deployment pipelines weirdly fascinating.
I am a B.Tech Computer Science graduate specializing in cloud infrastructure automation, Infrastructure as Code with Terraform and CloudFormation, and event-driven serverless backends on AWS.
I spend most of my time writing modular Terraform, troubleshooting IAM permission boundaries, and automating CI/CD pipelines so changes move from a git commit to an active AWS environment with zero manual intervention.
I test failure paths deliberately. On my multi-tier setups, I run kill-instance fault injection drills to verify Auto Scaling self-heals, set automated CloudWatch billing caps to prevent unexpected AWS charges, and enforce least-privilege IAM policies with zero wildcard permissions.
Automate before scaling
Manual deployments do not scale. Everything must be defined as code in Terraform or CloudFormation before adding traffic.
Measure before trusting
Opinions do not matter in infrastructure. P99 latency, error rates, and CloudWatch metrics dictate deployment decisions.
Simplify before adding
The best infrastructure is the one you do not have to manage. Prefer serverless Lambda and DynamoDB over heavy VMs.