aml-screening-service

This is a simple aml-screening-service

The Background

Banks are responsible for stopping money laundering. Hence it is necessary to avoid money transfers to terrorists and criminals. The list of such persons is published by EU:
http://eeas.europa.eu/cfsp/sanctions/consol-list/index_en.htm

Task

Your task is to implement simple webservice to compare given name against sanctioned list to detect such transfers.
Your API should have following functionality:

• Webservice with name as input and it should return if name match exists. True/false and extra data if needed.
• The webservice should be able to handle noisewords (example: the, to, an, mrs, mr, and) that need to be ignored.
• The webservice should detect names with spelling errors and also return as few "false positive" matches as possible.
• Webservices to add, update and remove sanctioned names.

Technical requirements

Implement using Java or Kotlin and Spring framework.
Sanctioned names should be stored in a database. (In-memory database can be used)
Open source libraries can be used.
Source code should be stored in a git based version control system.

Example

There is Sanctioned name: "Osama Bin Laden". Out of the box your API should detect at least the following user entries:

• "Osama Laden"
• "Osama Bin Laden"
• "Bin Laden, Osama"
• "Laden Osama Bin"
• "to the osama bin laden"
• "osama and bin laden"

In reality, name matching is not a trivial task. Matches could happen also based on:

• Substring matches - sl: "Robert", user: "Bert"
• Ignoring spelling errors - sl: "Madis", user: "Madus"
• Abbreviations - sl: "Joe Luis Webb", user: "Joe L. Webb"
• Noisewords - sl: "Mr. John Smith" user: "John Smith"

Instructions

This test can be either completed at home or during a pair programming exercise.

When implementing at home:
You should aim to spend no longer than 4 hours on your solution.
We would rather you complete less functionality to a good standard than rushing to try to complete it all.

Show us your workings by initialising a git repo and committing your changes as you go.
Your submission should be considered the quality you would expect to open a PR, not to merge one. We don't expect the submission to be production standard or gold plated.

Submit your sample by bundling your git repo (git bundle create name.bundle --all) and emailing the bundle to us (via your recruiter if you have one).
The submission will be followed up with a review call where we will ask you to walk us through your solution. We will then ask you to extend it with some additional functionality.

When implementing in a Pairing session:
We want to mimic real life as much as possible. This means that candidate should be able to google or ask questions.
We will spend 5 minutes introducing the challenge, 40 minutes coding and 15 minutes discussing the solution.

How to run

Install PostgreSQL

$ brew install postgresql

Start PostgreSQL service

$ brew services start postgresql # or "brew services run postgresql" to have it not restart at boot time

Run AmlScreeningServiceApplication with VM Option

-Dspring.profiles.active=development

Run with Docker

$ docker compose up

This should run postgreSQL and service in docker using the docker file

Service design and planning notes

Service component and data flow

Fuzzy Lookup from DB Logic