AI Resume Analyzer — CI/CD Pipeline with Docker and AWS

A serverless AI-powered resume analyzer built on AWS, with a full 3-stage CI/CD pipeline using GitHub Actions, Docker, and Amazon ECR.

Every push to main automatically runs tests, builds a Docker container, pushes it to ECR, and deploys it to AWS Lambda.

---

What It Does

Paste your resume and a job description. The app uses Amazon Bedrock (Claude) to return:

• Match score (0-100)
• Strengths
• Skill gaps
• Improvement suggestions
• Verdict (Strong Match, Good Match, Partial Match, or Not a Match)

Live demo: irbcho.github.io/AI-Resume-Analyzer

---

Architecture

GitHub Push
    |
    v
GitHub Actions CI/CD Pipeline
    |
    |-- Job 1: Run Tests (pytest)
    |
    |-- Job 2: Build Docker Image --> Push to Amazon ECR
    |
    |-- Job 3: Deploy Container Image --> AWS Lambda
                                              |
                                        API Gateway
                                              |
                                     Amazon Bedrock (Claude)

---

Tech Stack

Layer	Technology
AI / LLM	Amazon Bedrock (Claude Sonnet)
Compute	AWS Lambda (Container Image)
Container Registry	Amazon ECR
API	Amazon API Gateway
CI/CD	GitHub Actions
Containerization	Docker
Testing	pytest
IaC	AWS IAM, ECR, Lambda via CLI

---

CI/CD Pipeline

The pipeline has 3 jobs that run in sequence on every push to main.

Job 1 — Run Tests

Installs dependencies and runs 5 unit tests using pytest. If any test fails, the pipeline stops and nothing gets deployed.

Tests cover:

• CORS preflight (OPTIONS) request handling
• Missing resume input validation
• Missing job description input validation
• Invalid JSON body handling
• Full valid request with mocked Bedrock response

Job 2 — Build and Push to ECR

Builds a Docker image using the AWS Lambda Python 3.12 base image, tags it with the Git commit SHA for traceability, and pushes it to Amazon ECR. Also tags the image as latest.

Job 3 — Deploy to Lambda

Updates the Lambda function to run the new container image from ECR, waits for the update to complete, and verifies the deployment.

---

Screenshots

Pipeline Running — All 3 Jobs Green

All Workflow Runs Passing

Docker Image in Amazon ECR

Lambda Function — Package Type: Image

---

Project Structure

AI-Resume-Analyzer/
├── .github/
│   └── workflows/
│       └── deploy.yml        # CI/CD pipeline definition
├── tests/
│   └── test_lambda.py        # Unit tests
├── Dockerfile                # Container image definition
├── lambda_function.py        # Lambda handler
├── requirements.txt          # Python dependencies
├── index.html                # Frontend
└── README.md

---

How to Set This Up

Prerequisites

• AWS account
• GitHub repository
• Docker installed locally
• AWS CLI configured

Step 1 — Create ECR Repository

Go to AWS Console → ECR → Create repository.

Set the name to resume-analyzer and leave all other settings as default.

Copy the repository URI. You will need it in Step 4.

Step 2 — Create IAM User for GitHub Actions

Go to AWS Console → IAM → Users → Create user.

Name it GitHubActionsDeployer and attach these policies:

• AmazonEC2ContainerRegistryPowerUser
• AWSLambda_FullAccess

Generate an access key and copy the Access Key ID and Secret Access Key.

Step 3 — Add GitHub Secrets

Go to your GitHub repository → Settings → Secrets and variables → Actions → New repository secret.

Add these two secrets:

• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY

Step 4 — Create Lambda Function as Container Image

Go to AWS Console → Lambda → Create function → Container image.

Set the function name to ResumeAnalyzer and browse ECR to select your resume-analyzer repository. After creating, go to Configuration → Permissions and attach an IAM role with Bedrock access.

Step 5 — Push to GitHub

git add .
git commit -m "Initial CI/CD pipeline setup"
git push origin main

Go to the Actions tab and watch all 3 jobs turn green.

---

What I Learned

• How to containerize a Lambda function using Docker and the AWS Lambda Python base image
• How to push container images to Amazon ECR using GitHub Actions
• How to write unit tests with pytest and mock AWS SDK calls
• How to structure a multi-job CI/CD pipeline where each stage depends on the previous one
• How to manage IAM permissions for automated deployments
• The difference between Lambda Zip deployment and Container Image deployment

---

Author

Irene Cho — LinkedIn | Portfolio | GitHub