CortexData

What it is

The product

CortexData is a modular lending OS for Indian banks, co-operative banks, and NBFCs. Five composable modules — LOS, LMS, Co-Lending, PTC, CLO — eight loan products on day one, every customer-facing artefact rendered equally in English, ಕನ್ನಡ, and हिंदी.

The problem

What we set out to solve

India has 1,500+ co-operative banks, 9,500+ NBFCs, and a dozen Small Finance Banks. They run a patchwork of decade-old core banking systems, manual KYC, paper-based PTC issuance, and spreadsheet co-lending reconciliation. RBI Master Directions land every quarter; most platforms catch up two quarters later. The result: regulatory drag, recoverable losses, and lending teams that spend more time on compliance than on lending.

The build

How it actually works

We engineered CortexData ground-up against the RBI Master Directions corpus — IRACP, Digital Lending Guidelines, Co-Lending Model, Securitisation, Gold Loan, Priority Sector. Every regulatory return ships built-in: KFS, DSB, OSS-3, PSL Form A, IRACP. The ML decisioning stack is a 5-model fraud ensemble plus a calibrated PD model with full feature attribution — defensible to RBI, the Banking Ombudsman, and your credit committee.

Capabilities

• 8 loan products on day one — Personal, Home, Vehicle, MSME, Education, CC/OD, LAP, Gold
• 5 composable modules — LOS, LMS, Co-Lending, PTC, CLO
• Every RBI return shipped built-in — IRACP, DSB, OSS-3, PSL, KFS
• 5-model fraud ensemble with per-decision feature attribution
• Calibrated probability-of-default model with isotonic calibration
• EN · ಕನ್ನಡ · हिंदी across every customer artefact
• CBS-agnostic adapter pattern — Finacle, Flexcube, BaNCS, native
• Audit-immutable: SHA-256 hash chain, 10-year retention as standard