Hi there! And welcome to the technical test stage of the interview process for becoming a Kaizen Backend Software Engineer!

Normally at this point, you should already be aware of what will be expected of you over the next few hours.

So let's jump straight in!

We want you to create the REST API and backend resources of a web application that will be used as an "internal blog". The objective of the platform is to let Kaizen employees share updates and add comments to those updates through a single-page web UI, as shown below:

Users don't have to be authenticated. Anyone within the organisation can access this webapp and post updates and comments. They just have to enter their name and the message they want to post. (And optionally an image for updates if they want to)

Users can also delete comments. Again, it doesn't matter whether they posted the comment they want to delete or not, here any user is able to delete any comment, there is no concept of authentication or ownership. However, what we'll want in a situation where a comment is deleted, is to send an email to alert administrators about it.

Below is the full list of requirements. We recommend you to start by documenting the technical architecture you plan on implementing before doing any actual work. (Last section of this document)

• Amazon Certificate Manager (ACM)
• API Gateway V2
• Cloud9
• CloudFormation
• CloudWatch
• DynamoDB
• Identity and Access Management (IAM)
• Key Management Service (KMS)
• Lambda
• Secrets Manager
• Security Token Service (STS)
• Simple Email Service (SES)
• Simple Notification Service (SNS)
• Simple Queue Service (SQS)
• Simple Storage Service (S3)
• States (Step Functions)
• Systems Manager (SSM)
• Web Application Firewall (WAF)

• Cannot modify your IAM user or any already existing role/policy
• Cannot use or set up SSO

The architecture is as follows:

Two dynamodb tables are used and are indexed by the created_at field. The images will be uploaded to a separate endpoint that will store them in an s3 bucket.

When a comment is deleted, an SNS message will be sent. A lambda function subscribed to the Topic will receive the message and send an email to the administrator.

For IP whitelisting, a resource policy as follows will be used:

{
  "Effect": "Deny",
  "Principal": "*",
  "Action": "execute-api:Invoke",
  "Resource": "execute-api:/*/*/*",
  "Condition": {
    "NotIpAddress": {
      "aws:SourceIp": ["sourceIpOrCIDRBlock", "sourceIpOrCIDRBlock"]
       }
    }
}

For lamda auth IAM auth will be used .

1. Install dependencies:

$ poetry install

2. Setup pre-commit hooks before committing:

$ poetry run pre-commit install

$ poetry run pytest

This job makes extensive use of dependency injection. I know it is not very common in Python but nevertheless I think it is a great idea, and I love it. For this purpose I will use the kink library

For data input and data validation I use marshmallow combined with dataclasses. This way we ensure that all input data is valid and that the classes are consistent.

All tests are performed with `pytest. First I develop the whole application locally and test it with tests. Once everything is working, then I deploy to AWS.

I use pre-commit for development along with flake8, black and isort. I also use mypy for typing hints.