NEWDr FRIDAY, a wellness app we built and run on Amazon Web Services (AWS), is now on Google Play. View on Google Play →

← Case studies

How DrFRIDAY × Atrea launched a wellness screening app on Google Play in about five weeks, built on Amazon Web Services (AWS)

Download the case study (PDF, 2 pages)View on Google Play →

Summary

DrFRIDAY × Atrea, a doctor-led health group in Bengaluru, India, wanted to offer people a quick wellness check from the camera of their own phone. CloudNestle built the Android app and its serverless backend on Amazon Web Services (AWS). The app went live on Google Play 37 days after the scope was agreed.

Customer profile

DrFRIDAY × Atrea is a doctor-led healthcare collaboration in Bengaluru, India.

The challenge

The group wanted a patient to sign in with a one-time code, take a 30-second face scan, read the results in plain words and keep a history. A clinic needed to add scan credits for patients, and patients needed to buy more by themselves.

Dr FRIDAY is a wellness tool and does not give medical advice. It handles health-related data of people in India. Four constraints shaped the work:

  • India's Digital Personal Data Protection Act asks for clear consent and a way to delete an account.
  • Google Play's payment rules do not allow an app to send people to another payment method for in-app credits, so online purchase had to be built outside the app.
  • The face-scan service needs a licence from its provider. The licence was reactivated 16 days after the scope was agreed.
  • The scope was fixed, the plan was 34 days, and the team was one engineer.

The solution

CloudNestle built a serverless backend in the AWS Mumbai Region. Amazon API Gateway receives each request, and an AWS Lambda function checks the phone sign-in token before any data is read. One AWS Lambda function handles each task, with its own narrow permissions.

Amazon DynamoDB holds six tables. Each is encrypted with a customer-managed AWS Key Management Service (AWS KMS) key with automatic rotation, and each has point-in-time recovery. Consent is recorded on the server with the version of the Terms and a time stamp, and people accept updated Terms once before they continue.

Because Google Play does not allow in-app links to another payment method, credits are bought on a web page hosted on Amazon CloudFront and Amazon S3. A payment is credited exactly once, and an hourly job compares the payment provider's records with ours. Results can be sent to a person's WhatsApp through an Amazon SQS queue, if the person opts in.

The face scan is done by a specialist service outside the AWS account, and the Terms tell people which data goes to it.

Monitoring was built in from the start. Amazon CloudWatch has more than 40 alarms that notify the team by e-mail through Amazon SNS. An operations page shows which version runs in each of four environments, and a daily health e-mail summarises the day. An audit archive in Amazon S3 uses Object Lock in governance mode for 731 days. The whole setup is defined as code with the AWS Cloud Development Kit (AWS CDK).

How Dr FRIDAY is built on AWS: phone app, Amazon API Gateway, AWS Lambda, Amazon DynamoDB, Amazon SQS, Amazon EventBridge, Amazon CloudFront and Amazon S3, Amazon CloudWatch, audit log

Diagram by CloudNestle. Dashed boxes are outside our AWS account.

Results

  • Time to launch: live on Google Play 37 days after the agreed scope (23 September to 30 October 2026). The app was submitted to Google on day 32 of a 34-day plan.
  • Second release: version 1.1.0, with opt-in WhatsApp results and online credit purchase, went live on 10 November 2026.
  • Stability: no server errors (HTTP 5xx) on the production API from 10 to 20 November 2026. This counts errors at the API Gateway level.
  • Data protection set-up: all six production tables use customer-managed keys and point-in-time recovery; more than 40 production alarms were all in the OK state on 19 November 2026.
  • Quality: over 2,000 automated backend and infrastructure tests. Each change is tested, then deliberately broken to prove that the tests catch the fault.
  • A fair account of the schedule: three milestones finished early and two late. The scan reached the client's test environment 10 days after the plan, because of the licence wait and about five days of planned leave. Going live was 3 days after the plan, because of Google's review time.

About CloudNestle

CloudNestle Consulting & Services designs, builds and runs serverless applications on Amazon Web Services (AWS) from Bengaluru, India.

Dr FRIDAY is a wellness tool and does not give medical advice.

Talk to us about your project →