Our workMoney Infrastructure & AI Credit Risk

Buckyy

Live · Private accessVisit site

Payments, lending, and an AI credit engine built for markets traditional credit bureaus don't reach.

Buckyy's financial infrastructure — access is limited.

  • Next.js
  • Flutter
  • Python / FastAPI
  • Machine Learning

Overview

Buckyy set out to build money infrastructure for the internet — digital payments, Buy Now Pay Later, and intelligent lending, aimed at underserved financial markets where formal credit access is limited and existing scoring infrastructure is thin. The company's own positioning captures the ambition directly: “smart, simple, and tailored to you.”

We built the public-facing presence — website and mobile app — and then built the engineering core of the business: RINI, Buckyy's credit risk platform, which deserves its own section below because it's not a feature, it's the product.

Website & mobile app

Surfaces that make the pitch credible

Website (Next.js) — Buckyy's public site, explaining the platform's vision (money infrastructure, BNPL, credit scoring) and directing prospective institutional clients toward becoming a client.

Mobile app (Flutter) — the consumer-facing payments experience, including payment confirmation flows and the app shell for Buckyy's financial services.

These surfaces exist to do one job: make Buckyy's pitch to banks, NBFCs, and MFIs clear and credible, and give end users a working payments experience on top of it.

The core of the engagement

RINI — Credit Risk Automation Platform

Not a feature bolted onto Buckyy — the engineering core we built at the center of their business.

What RINI does

Score borrowers in seconds — not days

RINI is Buckyy's credit risk platform — a real-time scoring engine that evaluates borrower risk in seconds instead of days. In markets like Bangladesh, where formal credit history is thin or missing for most people, conventional bureau-dependent lending simply can't serve most applicants. RINI was built to solve that: score borrowers reliably using the data that's actually available, not the data a traditional system assumes exists.

A borrower applies through a lender's system, RINI evaluates the application, and the lender receives a decision score, a risk classification, and a clear explanation of what drove it — fast enough to fit inside a live application flow, not a batch process.

The engine

Accurate enough to trust with real lending decisions

The scoring engine we built for Buckyy was refined and validated against a large historical lending dataset. On that validation set it reached 97.81% accuracy, with recall, precision, and F1 all above 98%. The exact modeling approach, feature set, and architecture stay with Buckyy — what matters here is the result: a scoring engine accurate enough to trust with real lending decisions at scale.

97.81%

Accuracy

98.95%

Recall

98.31%

Precision

93.39%

Specificity

98.63%

F1 score

Confusion matrix

Validation results — correct predictions highlighted.

CorrectMiss
Predicted FailurePredicted Success
Actual Failure
21,044
1,489
Actual Success
921
86,498

Thin-file markets

Built for markets where credit history is thin

Rather than relying on the kind of formal credit bureau data that barely exists for most Bangladeshi borrowers, RINI draws signal from a broader picture of financial behavior — enough to score first-time and thin-file applicants that a conventional system would simply turn away for lack of data.

Explainable by design

Explainable by design, not as an afterthought

Every score RINI produces comes with a clear, human-readable explanation of what drove it. In a regulated lending environment, a score nobody can justify is a score nobody can use — so explainability isn't a reporting feature bolted on afterward, it's built into how RINI produces a decision in the first place.

Compliance

Compliant and audit-ready

RINI was designed from day one against Bangladesh Bank's regulatory framework for credit scoring — consent-based data handling, encrypted data at rest and in transit, full audit trails, and active bias monitoring across borrower segments. It's built to the bar required for formal credit bureau licensing, not just internal use.

Deployment

Deployment flexibility

RINI adapts to an institution's technical maturity — from direct API integration into an existing loan origination system, to a standalone hosted dashboard for lenders without in-house risk infrastructure.

Outcome

Live · Private access

RINI — the credit risk platform we helped develop for Buckyy — is live and in active use. Access is permissioned directly to partner institutions rather than public, consistent with its role as Buckyy's financial infrastructure rather than a consumer product.

Visit buckyy.com

Readyto startthe project?