Cloud & DevOps Engineer building reproducible AWS infrastructure with Terraform and GitHub Actions. Focused on least-privilege IAM, automated S3/CloudFront static delivery, and event-driven serverless backends.
Counted across 3 projects
DuoKart, Dorm-Dish, Voting. All infra as code
Cut via GitHub Actions caching on the voting-system pipeline
Map before
you touch
Scope the system and blast radius first. Every later step stays cheap because this one was honest.
IaC
Everything version-controlled and reproducible from a fresh account.
CI/CD
Cached pipelines with least-privilege deploy roles.
Observe
Alarms that page on symptoms, dashboards for the rest.
Secure + Cost
IAM least-privilege by default, $0 idle serverless where it fits.
“Good infrastructure is invisible: version-controlled, reproducible, and least-privilege by default.”
Survived a kill-1-EC2 drill with zero dropped responses

Multi-tier e-commerce cloud infrastructure across 2 AZs on AWS us-east-2. Internet-facing ALB routing to private Auto Scaling EC2 instances, Multi-AZ RDS MySQL, and SQS/DLQ decoupled event-driven Lambda order workers.
Role: Cloud Infra Engineer, team of 2
Evidence: ASG auto-healed in 3.5 min, DLQ redrive after 3 failures
Failure learned: Monoliths drop orders when compute dies mid-request; fixed with SQS decoupling plus DLQ redrive
$15+/mo idle cut to $0.00/mo

Serverless student accommodation platform on AWS: Cognito auth, Lambda + API Gateway backend, DynamoDB multi-table design, S3 media storage, and CloudFront global CDN distribution. Scaled down to $0/mo idle cost.
Role: Cloud and Backend Engineer
Evidence: CFN teardown, CloudFront edge, 350ms cold starts
Failure learned: Idle EC2 billed through academic breaks; fixed by going serverless
500+ students voted with zero downtime

React + Amplify frontend with Cognito multi-role auth. 6 Python Lambda REST microservices, DynamoDB On-Demand for burst vote writes, S3 audit export, and GitHub Actions CI/CD cutting deploy time to 3 minutes.
Role: DevOps and Backend Engineer
Evidence: 12m to 3m deploys, 180ms P99 latency
Failure learned: Paper tallying caused errors and slow results; fixed with DynamoDB burst writes plus S3 audit export
Crashed units diagnosed in under 30 seconds
Modular Linux system administration and automation toolkit in Bash. Features systemd service supervision with auto-restart, sub-30s network diagnostics, POSIX file permission auditing, and ShellCheck CI linting.
Role: Sole builder, Bash
Evidence: systemd supervision, ShellCheck CI, dry-run validation
Failure learned: Hand-checking units and sockets wasted on-call time; fixed with strict scripts
STOP BABYSITTING DEPLOYS.
Architecture review, CI/CD rescue, or full DevOps setup. I reply in 24h IST with a plan, not a pitch.