Employee JSON Converter — Home Assignment

Thank you for taking the time to work on this assignment. The goal isn't to catch you out — it's to give us a concrete piece of code to discuss together afterwards. We care far more about how you reason about an unfamiliar codebase, the trade-offs you weigh, and the quality bar you hold than about how many boxes you tick.

The project

This is a small Spring Boot service (json-converter). Its job: load employee data from several data sources, each of which returns JSON in a different shape, normalize every record into a single common Employee model, and combine the results.

The moving parts:

• DataSourceDAO implementations — each represents one external source and returns a raw JSON string. Different sources use different JSON structures (a flat array, a keyed dictionary, a nested object…).
• JsonEmployeeMapService implementations — each knows how to parse one structure into Employees.
• MapRouterService — inspects an incoming JSON payload and routes it to the right mapper.
• EmployeeDataLoadService — pulls from all sources and aggregates the normalized employees.

Take some time to read the code and the tests until you're comfortable with how it fits together.

Getting started

Requirements: JDK 21 and the bundled Maven wrapper (no global Maven needed).

# Run the test suite
./mvnw test

# Run the application
./mvnw spring-boot:run

What we'd like you to do

Part 1 — Refactor & harden

Bring this code up to the standard you'd expect of production code you own. We've intentionally left it rough in places. We won't list what to change — finding that is part of the exercise — but in general we care about correctness, reliability, readability, and maintainability.

A good first step is to build the project and run the tests, and treat what you find as your starting point.

Part 2 — Add a new data source structure

A new source needs to be integrated. It returns this structure:

{
  "results": [
    {
      "employeeNumber": "W-2001",
      "personal": { "fullName": "Grace Hopper" },
      "work": { "emailAddress": "ghopper@navy.mil", "position": "Rear Admiral" },
      "employmentStatus": "ACTIVE"
    },
    {
      "employeeNumber": "W-2002",
      "personal": { "fullName": "Katherine Johnson" },
      "work": { "emailAddress": "kjohnson@navy.mil", "position": "Mathematician" },
      "employmentStatus": "TERMINATED"
    }
  ]
}

Mapping rules:

Source field	Employee field	Notes
employeeNumber	id
personal.fullName	fullName
work.emailAddress	email
work.position	role
employmentStatus	active	ACTIVE ⇒ true; anything else ⇒ false

Integrate this end to end so it flows through the existing pipeline alongside the current sources, the way a real new source would. How easy or hard this is should tell you something about Part 1.

Ground rules

• Java 21, Maven, keep the project building and the test suite green.
• Add or change tests as you see fit — we value tests.
• You may restructure code, rename things, and adjust interfaces if you can justify it. Avoid pulling in heavy new dependencies without a good reason.
• Time-box it to roughly 3–5 hours. This is deliberately more than you may finish perfectly — we'd rather see a few things done well, with clear notes on the rest, than everything done hastily.

What to submit

1. Your code — as a fork/branch with a pull request, or a zipped repository.
2. A short NOTES.md containing:
   • the issues you found and how you addressed them (or would, with more time);
   • the key decisions and trade-offs you made, and why;
   • anything you deliberately left out and what you'd do next.

The NOTES.md matters as much as the code — it's the basis for our follow-up conversation, where we'll ask you to walk us through your reasoning.

A note on AI tools

Use whatever tools you'd normally use, AI assistants included. The only thing we ask is that you fully understand and can defend every decision and every line in the follow-up interview — we'll go deep.