CortexData
Platform · Architecture

Engineered for Indian banking realities.

CBS-agnostic. Multi-tenant. Air-gap-ready. Built so your operations team can run it.

CortexData runs as your full-stack lending OS or as a layer over your existing CBS. The same platform code deploys on AWS Mumbai, your private data centre, or air-gapped on-prem. We optimised for the way Indian banks actually run technology — not the way Silicon Valley vendors think they should.

  • TypeScript microservices on Postgres + Prisma; Python for ML
  • CBS-agnostic adapter port: Finacle, Flexcube, BaNCS, or native ledger
  • Multi-tenant from day one — soft or hard isolation
  • AWS ap-south-1 with Terraform, or air-gapped on-prem K8s
  • Prometheus + Grafana observability, alert rules included
  • RBI data-localisation by design — data never leaves India
Platform · Architecture module · highlights
  • 01
    TypeScript microservices on Postgres + Prisma; Python for ML
  • 02
    CBS-agnostic adapter port: Finacle, Flexcube, BaNCS, or native ledger
  • 03
    Multi-tenant from day one — soft or hard isolation
  • 04
    AWS ap-south-1 with Terraform, or air-gapped on-prem K8s
Architecture

The shape of the platform.

Each capability below is real architecture, not a marketing promise. We'll walk you through the deployment topology and show you the running stack on a discovery call.

CBS-agnostic adapter port

Same business logic against Finacle, Flexcube, TCS BaNCS, or a CortexData-native double-entry ledger. Switch CBS without touching origination or LMS code.

Microservices boundaries

API Gateway, LOS, LMS, CLO, Co-Lending, PTC, KYC, CBS Integration, Reports, Notifications. Independent deploy + scale per service.

Postgres + Prisma seam pattern

Postgres for OLTP. Every service exposes a narrow Prisma seam port — testable in-memory in CI, swapped to real DB in production. 12 production seam adapters today.

Multi-tenant from day one

Per-tenant configuration: PSL targets, sanction matrix tiers, exposure caps, languages, KYC vendors. Shared compliance core. One platform, many institutions.

AWS Mumbai (ap-south-1) by default

Pre-built Terraform for AWS ap-south-1. Multi-AZ Postgres, S3 Object Lock for audit retention, KMS-managed secrets, ALB + WAF, VPC with isolated subnets.

Or run on your hardware

Same Kubernetes manifests work on private cloud, VMware, or bare-metal. We've shipped to UCBs running their own data centres in tier-2 cities.

Observability built-in

Prometheus metrics on every state transition. Alert rules for stuck queues, NPA breaches, reconciliation gaps. Grafana dashboard JSON ships with the platform.

Security posture

OWASP-aligned controls, secrets in KMS / HashiCorp Vault, mTLS between services, RBAC + maker-checker at the application layer, VAPT-tested.

Data locality + encryption

RBI data-localisation by design — all customer data stays in India (ap-south-1 or on-prem). At-rest encryption (KMS), in-transit (TLS 1.3), field-level for PII.

EN · ಕನ್ನಡ · हिंदी

Multi-language as a first-class architectural concern: locale resolution at request boundary, ICU-based formatters for ₹ + Indian dates, three-language artefact rendering.

Deployment topologies

Three ways to run CortexData.

AWS Mumbai (ap-south-1)

Recommended for new deployments
  • Multi-AZ Postgres (RDS or Aurora)
  • S3 Object Lock for audit retention
  • ALB + WAF + Shield
  • KMS-managed secrets, VPC isolation
  • EKS cluster with managed node groups
  • Pre-built Terraform module

Private cloud / on-prem

For institutions with existing data centres
  • Kubernetes (vanilla or OpenShift)
  • Postgres (HA, your distribution)
  • MinIO for audit archive (WORM enabled)
  • Prometheus + Grafana on-cluster
  • Air-gap mode for fully-isolated networks
  • Helm charts + manifests provided

Hybrid

Common transition pattern
  • Origination on cloud, LMS on-prem
  • Or non-prod on cloud, prod on-prem
  • Cross-network mTLS service mesh
  • Audit aggregation to a central sink
  • Migration runbooks shipped
  • Full data-locality preserved

Frequently asked questions

Walk through the architecture with our team.

We'll show you the running deployment topology, the CBS adapters, the audit-chain wire-up, and the multi-tenant configuration on a 60-minute architecture deep-dive.