• Problem

Our client had already built a strong foundation with their core lending platform, but one critical piece was holding them back: the credit scoring module. At first glance, the system worked — it could evaluate borrowers and help decide whether to issue loans.

But beneath the surface, there was a problem.

Each new financial institution had its own scoring criteria. Adjusting the solution to them required developers to make changes directly in the code. That meant they were constantly pulled into repetitive tasks, tweaking rules for age, salary, loan history, or expenses, instead of focusing on innovation.

On top of that, monolithic architecture didn’t allow to introduce changes on the go.

This rigidity created a bottleneck. For our client’s customers, it meant slower onboarding and limited flexibility. For our client themselves, it meant higher costs, wasted resources, and a product that couldn’t scale as quickly as the market demanded.

• Our solution

We added the new logic to conveniently tweak credit scoring rules directly from the tool’s interface. This allowed our client’s staff to adapt them to each new customer without involving developers.

We also extended the app’s functionality to process smaller loans automatically. It collected various data on applicants from their profiles and other sources and applied a grading algorithm. For example:

1. Employment stability:

– Full‑time job with 3+ years at current employer → +15 points

– Frequent job changes in the past 2 years → –10 points

2. Savings & assets:

– Maintains a savings account with consistent deposits → +10 points

– No savings or assets → –10 points

3. Payment behavior:

– No late payments in the past 12 months → +15 points

– Multiple late payments in the past year → –20 points

4. Credit utilization:

– Uses less than 30% of available credit → +10 points

– Maxed out credit cards or overdrafts → –15 points

(the reference doesn’t illustrate the actual scoring criteria and algorithm)

Summarizing results, the tool worked out the final grade and made a decision on approving/rejecting the loan. Riskier or more complex loans still required manual review after the automatic assessment.

Pursuing our vision of tech efficiency, we also re‑engineered the platform’s architecture. We broke the monolithic application into microservices that made the system easier to support, customize, and connect with other tools.

• Workflow

The development team operated fully remotely, following Agile methodology with structured sprints and bi-weekly demos to ensure continuous feedback and alignment. Transparent communication and progress tracking were maintained through Jira for task management and Slack for daily collaboration.