12. Overview#
This is a quick take on risk segmentation. The objective is to break down credit risk into its fundamental components so that we can make use of more targeted ML systems, risk assessments and controls.
Risk segmentation also enables us to design modular and independent services (i.e. software systems) for each risk component, and balance/control overall risk exposure against “growth vs profit” objectives via a dedicate business layer, call it Credit Risk Orchestrator. This gives analysts, product specialists, compliance officers and others easy control and observability without needing to interact with the underlying complexity. It’s the magic of abstraction; if done well, it’s a super-power. Cherry on top is enabling risk assessments and controls to be tweaked via an intuitive UI with role-based access control and approval rights.
Of course, all of this needs a lot of work and this is just a case study. So I dedicated a couple of hours to brainstorm on the topic and created some brain dumps. They should cover the following topics: