Reliable financial modelling system calculations pose a certain level of complexity. Consequently, we have partitioned the system into hundreds of independent modules that handle different aspects of the calculations.

This approach ensures easy maintenance, testability and extensibility. Building the architecture of a system that models thousands of economic scenarios spread over decades of data was a tremendous challenge.

To fully comprehend the monumental gravity of this task you have to add the factors - performance, usage and assets under management, in delivering the optimum solutions to our client's needs.

We had to engineer a solution to reflect the multi-tier taxation laws across geographical locals stipulated by regulatory frameworks. The scope of consideration was not purely limited to country, municipality and city regulations, but to the way business is conducted in that area (e.g. some operators and markets would pay the tax that the independent farmer is due in their name, in order to reduce the administrative effort their partners are facing).

Due to the project being an organisation-wide system, we had to create a unified design to enable cross-country collaboration to become frictionless. The end product was a system that allows consistency in terms of usage, easy modification and testing of existing taxation logic (e.g. to reflect new laws/rules and amend existing ones).

We carried out an extensive overhaul consisting of changes to table structure, indexes, adding partitioning and the core logic of how data is inserted or read. This resulted in almost 300% system performance increase and no undesirable impact on the existing structure that we had to reconfigure.