AWS Builder Center
From Architecture to Deployment — Building the Image Editing App with AI for Bharat

From Architecture to Deployment — Building the Image Editing App with AI for Bharat

A hands-on technical walkthrough of implementing the Image Editing App under AI for Bharat — an AWS initiative in India that promotes accessible, industry-aligned AI and cloud learning. This article covers the complete build journey from architecture understanding to authentication, backend logic, API integration, frontend deployment, and testing, while reflecting on practical engineering insights and community-orientated learning impact.

Workflow — Understanding the Application Journey

The Image Editing App follows a cloud-native serverless workflow, where each AWS service contributes to a well-defined application path.
The request lifecycle moves from authentication to backend execution and finally returns output to the frontend interface, ensuring secure, scalable, and predictable behaviour.
User → Cognito Authentication → API Gateway → AWS Lambda → DynamoDB → Response to UI

This workshop helped me understand how these services operate together as a distributed, event-driven system, rather than functioning as isolated components — a key mindset for modern application design.

Create Amazon Cognito User Pool — Authentication First

We began by configuring Amazon Cognito, creating a User Pool and adding a user account.
In this application, Cognito served as the entry-point security layer, ensuring users authenticate before accessing application workflows.
This reinforced an important architectural principle:
Authentication must be established before functionality.
Amazon Cognito User Pool

It aligns with secure application development practices, especially when applications are designed for public use.

Create DynamoDB Table — Defining the Data Layer

Next, we configured an Amazon DynamoDB table to support application-level storage.
This stage strengthened my understanding of:
lightweight and scalable storage design
stateless execution behaviour
serverless-ready data handling
DynamoDB integrates efficiently with Lambda, enabling fast, event-driven data workflows without additional infrastructure overhead.

Create AWS Lambda Function — Backend Execution Logic

The AWS Lambda function worked as the processing engine of the application.
Here, I implemented and understood:
request parsing and validation
interaction with DynamoDB
structured response handling
This deepened my exposure to event-triggered compute execution, where code runs only when invoked — encouraging efficiency, modularity, and operational discipline.

Create Amazon API Gateway — Connecting UI to Logic

Once backend logic was ready, we exposed it through Amazon API Gateway.
This enabled:
secure routing between UI and backend
method integration mapping
consistent request-response flow
API Gateway made it easier to visualise how application layers communicate, reinforcing the importance of API-first application design.
It also strengthened concepts around:
  • modularity
  • integration flexibility
  • maintainability at scale

Create AWS Amplify Application — Deploying the Frontend

The frontend was deployed using AWS Amplify, which provided:
  • seamless hosting workflow
  • build + deploy integration
  • environment consistency
User Sign-in Interface

Seeing the app function end-to-end transformed the architecture into a working, testable system.

Test the Application — Validating Behaviour

Finally, we validated the complete application by DynamoDB logging.
DynamoDB Logging

This phase was not just about functionality — it helped me analyse:
  • response behaviours
  • dependency flow
  • traceability across services
It encouraged thinking from both an engineering perspective and a product-execution lens.

Result — Application Execution & Prompt Workflow

After completing the testing and validating the application behaviour, I tested the image editing and AI-driven enhancement flow using the prompt interface.
  • Upload Sample Image
Sample Image

  • Select Target Area → Highlighted the car region in the image.
Selected the Car

The Debug Console displayed metadata and processing details.
  • Enter Prompt → Used the following prompt:
Make the driveway clear and empty.

  • Generate Output → the application produced two AI-generated edited images.
Output Images


Taking It Forward

This workshop strengthened my understanding of:
authentication, backend workflows, APIs, deployment, and testing — and I now look forward to delivering these learnings to rural and student communities through my initiative:
“For the Students, By the Students”
Because talent exists everywhere, opportunity must too.


I will keep sharing more learnings and implementation experiences from upcoming AI for Bharat workshops and related initiatives.
Stay Connected — Grow Together 🤝  Bhoomi! 

References:

Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article