Architecture & Technology Stack

AI-Native Data Fabric Architecture powered by SciKIQ — Connect, Curate, Control, Consume methodology with 167+ pre-built connectors.

Architecture

SciKIQ AI-Native Data Fabric

A unified data fabric platform using the Connect-Curate-Control-Consume (4C) methodology — deployed in 30-45 days with 167+ pre-built connectors, zero-code interface, and embedded GenAI.

SciKIQ 4C Framework — Data Fabric Pipeline
Connect • Curate • Control • Consume
Data Sources
GCash Core
BancNet Switch
InstaPay / PESONet
Event Streams
SAP / ERP / APIs
1. Connect
Data Integration Hub
167+ connectors • Batch + Streaming
Auto-ingestion • Schema mapping
2. Curate
Data Prep Studio
GenAI enrichment • MDM • Dedup
Logical modeling • Orchestrator
3. Control
Governance Suite
Metadata catalog • Multi-hop lineage
RBAC • Policy enforcement • dQ
4. Consume
AI & Analytics
NLQ GenAI Studio • ML Studio
BI connectors • Data Products • APIs
SciKIQ Six Integrated Layers
AI-Native • Zero-Code • Cloud-Agnostic
Layer 1
Data Ingestion & Integration
167+ pre-built connectors for SAP, Oracle, cloud & legacy systems. Auto-ingestion with batch + streaming. AI-powered schema mapping.
Layer 2
Automated Governance & Semantics
RBAC, automated dQ checks, sensitivity tagging, raw-to-report lineage, semantic modeling for business KPIs.
Layer 3
Data Processing & Curation
Standardization, MDM, deduplication, GenAI-driven enrichment, real-time or scheduled processing.
Layer 4
Data Product Factory
Convert curated datasets into reusable, governed data products. Auto-generate APIs, dashboards & access rules.
Layer 5
Data Marketplace
Internal/external data product discovery. API-enabled access with usage controls. Monetization for partnerships.
Layer 6
Consumption & AI Layer
GenAI Studio (NLQ), ML Studio, AI Agents for autonomous workflows, BI connectors (Power BI, Tableau, Looker, Qlik).
Cloud Infrastructure

Azure Cloud Architecture & Environments

UDP hosted on Azure Landing Zone with 3 subscriptions — Dev, Test & Prod, enabling 4 data environments.

Environments
Dev
Test
Pre-Prod
Prod
On-premise connectivity maintained for hybrid data integration
6 Pillars of Requirement
Management & Governance
Networking & Security
Identity & Access
Logging & Monitoring
IaaS & PaaS Services
Applications & Workloads
Service Distribution

Azure Services — Business Unit View

Centralized storage with dedicated compute for each business unit. Data Management Zone handles all inbound/outbound traffic, Data Zone processes workloads.

Singtel Data Repository — Centralized Storage
Business Unit 1
Dedicated Compute
Business Unit 2
Dedicated Compute
Business Unit 3
Dedicated Compute
Azure Service Inventory
SciKIQ Data Fabric SciKIQ NLQ Studio SciKIQ Governance Storage Account Key Vault MS-SQL DB Postgres Redis Cosmos DB Airflow EventHub Azure Monitor Log Analytics Azure VM VNET/Subnet Azure Firewall Active Directory
Packaged Capabilities

Packaged Data Workspace (PDW)

Bundled business capabilities that enable seamless onboarding of new business units with all data workspace capabilities included.

Storage
Compute
ETL / ELT
Security
Monitoring
Access Ctrl
Infrastructure as Code

Unified IaC — CAFs Terraform for SciKIQ Data Fabric

Cloud Infrastructure Setup: Complexity & Dependencies
Vision to Reality

Cloud Infrastructure hosting requires multiple cross-domain teams to work on multiple tools & services with dependencies:

1. Too many cross-domain dependencies!
2. Too many Azure Services needed!
3. Too many automation tooling choices!
Cross Domain Teams
Cloud-INFRA Team
Network Team
Cloud DevOps Team
Data Team
Common Services
Firewalls IP Address allocations On-premise Networking Observability & monitoring DevOps / CICD
Azure Services
Management Group
Subscriptions
Azure Active Directory
Azure Virtual WAN
Storage Account (Blob/ADLS)
Azure VNET/Subnet
DNS
SciKIQ
Azure Firewall
Azure Monitor
Data Factory
Key Vault
Log Analytics
EventHub
Cosmos DB
Azure SQL DB
Postgres
Azure VM
Airflow
Automation Tools
Azure ARM
Azure Bicep
Terraform Chosen
CAFs Terraform Chosen
Ansible
Puppet
Chef
PowerShell
Python / GO
Azure CLI
Design Principles & Key Considerations
1Package Business Capabilities — Data-workspaces to enable with all capabilities, ability to onboard NEW BU seamlessly
2Metadata & Config Driven — Everything as Code or Config
3Framework — Ensure consistency, reusability, maintainability, readability
4Dependency Injection — Seamless integrations for cloud infra layers and teams
5Testing Driven — All cloud provisioning supported by LOCAL testing & automated tests
6Cloud Agnostic — Capability to scale to AWS as well, not limited to Azure
Key Patterns Applied
Level Dependency Injection & Key Association Patterns address team dependencies
Iterate on Everything pattern — collection of deployments worked well
Azure CAFs Framework provides a foundation to build upon
Declarative & maintainable — everything as configuration or code
AI
AI Analyst

Welcome to the Data Manager AI Analyst. I can help you with:

  • Revenue leakage analysis & recovery strategies
  • Fraud pattern detection insights
  • Compliance & AML risk assessment
  • Data quality recommendations
  • Executive-level briefings

Ask me anything about your banking data platform.