Portfolio projects that get interviews

Why projects matter

Hundreds of candidates apply for each fresher cloud role with the same certificate. A public GitHub repository with working infrastructure code and a clear README makes you different. In interviews, projects give you real stories: what you built, what broke and how you fixed it. Interviewers trust those stories more than definitions.

Project 1: static website with HTTPS

Host a personal site or résumé page in a private S3 bucket. Put Amazon CloudFront in front, using origin access control (OAC) so only CloudFront can read the bucket. Request a free certificate from AWS Certificate Manager in us-east-1 and attach it to CloudFront. If you own a domain, create a Route 53 alias record pointing to the distribution. What it proves: S3, CDN, TLS, DNS and bucket security. Cost: usually a few rupees a month plus the domain.

Project 2: serverless API

Build a small API, for example a to-do list or a library book tracker: API Gateway (HTTP API) receives requests, a Lambda function in Python or Node.js handles them, and DynamoDB stores the items in on-demand mode. Give the function an IAM role that can touch only that one table. Add CloudWatch logs and an alarm on errors. What it proves: serverless design, IAM least privilege, NoSQL data modelling, monitoring. Cost: often near zero at low traffic.

Project 3: three-tier app with Terraform

Write Terraform code for a custom VPC across two Availability Zones with public, private and database subnets. An Application Load Balancer sits in the public subnets, an Auto Scaling group of EC2 instances runs the app in private subnets, and RDS (Multi-AZ if your budget allows) sits in the database subnets. Security groups chain: internet to ALB on 443, ALB to app on the app port, app to database on 5432 or 3306. What it proves: networking, high availability, infrastructure as code, the designs asked about in SAA-C03. Cost warning: the NAT gateway, load balancer and RDS charge every hour. Build it, test it, screenshot it, and run terraform destroy the same day.

terraform init
terraform plan -out plan.tfplan
terraform apply plan.tfplan
# test, take screenshots, note the outputs
terraform destroy

Project 4: monitoring and cost dashboard

Build a CloudWatch dashboard for one of the projects above: request count, error rate, latency, Lambda duration, CPU of the Auto Scaling group. Add alarms that email you through SNS. Then add the cost side: an AWS Budget with alerts, cost allocation tags such as Project and Environment, and a Cost Explorer view grouped by tag. What it proves: you think about operations and money, which is exactly what employers worry about.

How to document a project

A README is the front page of your repository. Recruiters may read only this, so make it clear in two minutes.

# Serverless book tracker (AWS)

## What it does
A REST API to add, list and delete books. Live demo: (link, or a short screen recording)

## Architecture
(diagram: client -> API Gateway -> Lambda -> DynamoDB; CloudWatch logs and alarm)

## Services and why I chose them
API Gateway HTTP API (cheaper and simpler than REST API for this use), Lambda (no servers),
DynamoDB on-demand (pay per request, scales to zero).

## Security choices
Lambda role can only read and write the Books table. No access keys in code. HTTPS only.

## Cost estimate
About US$0 to 1 per month at 10,000 requests (AWS Pricing Calculator link).

## How to deploy and tear down
terraform apply ... / terraform destroy ...

## What I learned and what I would improve
Cold starts, DynamoDB key design, adding authentication with Cognito next.