Deep expertise in the systems that financial services runs on.
Three practice areas. All built around the problems that matter most to lenders, fintechs, and credit bureaus operating in APAC.
Decisioning & Scoring
Credit decisioning systems fail in predictable ways. Policies change without audit trails. Models go to production without challenger testing. Regulators ask for decision explanations and nobody can produce them. Scorecards drift from their original design because nobody documented the intent. And when something goes wrong, there's no rollback.
We've seen this pattern at national credit bureaus, at regional lenders, and at digital banks. The technology is rarely the root cause — it's the absence of governance, versioning, and explainability built into the system from the start.
Decisioning engine architecture and build
Design and implementation of rules-based and scorecard-driven decisioning systems — covering decision flows, decision trees, decision matrices, and scorecard integration. API-first, container-native, deployable on any cloud or on-premise environment. Business users can own and modify decision logic without engineering involvement.
Decision explainability
Every decision produces a structured explanation — which rules fired, which scorecard bands applied, what the contributing factors were. Built for regulatory audit, not as an afterthought.
Champion/challenger framework
Run parallel policy versions in production with controlled traffic splits. Measure outcomes. Promote the winner. Retire the loser. Continuous improvement with controlled credit risk.
Version control and policy diff
Full history of every policy change — whether it's a decision tree branch, a matrix threshold, or a scorecard weight. Side-by-side diff between any two versions. Know exactly what changed, when, and who approved it — before and after deployment.
Controlled deployment pipeline
Separate the configuration layer from the runtime. Policy changes go through a defined approval and staging process before they touch production. Rollback in minutes, not days.
Bureau data orchestration
Multi-bureau pull, response normalisation, caching, and fallback logic. Reduce bureau costs, improve hit rates, and insulate your decisioning layer from upstream data quality issues.
Alternative data scoring
Integration of non-traditional data sources — telco, transaction, behavioural — into scoring pipelines. Includes FICO standard input format compatibility where required.
Score delivery API
Clean, documented, versioned API layer for score and decision delivery to downstream systems. Includes monitoring, alerting, and SLA tracking.
Relevant engagements
- Architecture advisory for decisioning platform modernisation at Malaysia's leading credit bureau
- Decisioning workflow design and delivery for fraud monitoring platform targeting a Malaysian digital bank (pre-sales, live demo available)
- Alternative data scoring platform built during Experian engagement — led development of a telco-based credit scoring solution aggregating data from three Indonesian telcos, with scores consumed by leading Indonesian banks
Onboarding & Digital Identity
Digital onboarding is where financial institutions lose customers they've already acquired. Drop-off at identity verification. Friction at document upload. Manual review queues that take days. Fraud that slips through because the checks aren't connected. And compliance teams that can't see what happened at each step.
The underlying issue is almost always the same: onboarding was built as a sequence of point solutions — an eKYC vendor here, a document check there, a manual review step bolted on — rather than as an integrated workflow with a single audit trail.
End-to-end onboarding workflow
Full application flow from initial data capture through identity verification, document collection, fraud screening, credit assessment, and account opening. Built as a single integrated workflow, not a chain of disconnected steps.
eKYC and identity verification integration
Integration with leading identity and risk intelligence providers including Monnai (5B+ identities, 1,000+ signals) and Veriff (eKYC). Configurable verification levels based on risk tier.
Document collection and verification
Structured document upload, classification, and verification routing. Supports manual review queues with SLA tracking and escalation.
Fraud screening integration
Transaction fraud scoring (Kount), digital identity risk (Monnai), and mule account detection integrated into the onboarding decision flow.
Audit trail and compliance reporting
Every step of the onboarding journey is logged with timestamps, decision outcomes, and operator actions. Produces a complete audit trail for regulatory review.
Referral and escalation management
Applications that can't be auto-approved are routed to the right review queue with context, SLA timers, and escalation paths. No applications lost in the system.
Relevant engagements
- Fraud monitoring and onboarding platform for a Malaysian digital bank — integrating Monnai, Kount, Veriff, and CTOS bureau data (pre-sales, live demo available)
- Onboarding workflow design and demonstration for major financial institutions across APAC (Experian engagements)
Platform Modernisation & Data Engineering
Legacy financial systems don't fail suddenly — they accumulate risk. A core platform built for a different era, running on infrastructure nobody fully understands, with integrations that are undocumented and fragile. The business wants to move faster. The technology team knows that moving faster on this foundation is dangerous.
At the same time, most institutions are sitting on data they can't use. It arrives from multiple sources in inconsistent formats, nobody is confident in its quality, and by the time it reaches an analyst it's already stale. The analytics capability the business needs — credit performance monitoring, fraud trend detection, regulatory reporting — is blocked by a data infrastructure that was never designed for it.
The answer isn't always a full replacement. Sometimes it's adding a modern data layer on top of what exists. Sometimes it's rebuilding the processing pipeline with proper quality controls. Sometimes it's simply adding observability so you can finally see what's actually happening.
Architecture assessment and modernisation roadmap
A structured review of current systems, integrations, data flows, and team capability — producing a phased roadmap with risk-rated options. Not a "big bang" recommendation, but a realistic path from where you are to where you need to be.
Cloud-native data platform
Enterprise data lake design and implementation — built for credit, fraud, and regulatory analytics workloads. Tiered storage architecture (raw, validated, aggregated), optimised for both batch reporting and interactive query. Cloud-agnostic and container-native.
Configurable data ingestion platform
A reusable ingestion framework that handles data quality checks, business rules validation, and automatic trend shift detection out of the box. When a data source changes behaviour — a new null pattern, a distribution shift, a volume anomaly — the platform flags it before it reaches downstream models or reports.
Data analytics and performance monitoring
Dashboards and analytical pipelines for credit portfolio performance, fraud trend monitoring, operational KPIs, and regulatory reporting. Built so business teams can answer their own questions, not just the engineering team.
Microservices migration
Decomposition of monolithic systems into independently deployable services — with API design, service boundary definition, and data migration strategy. Container-native, deployable on any cloud or on-premise environment.
Observability and monitoring
Full-stack observability covering system health, data quality metrics, and business KPIs in a single view. Built so the business can see what's happening in real time, not just after an incident.
Report automation
Automated generation of structured reports — XML, PDF, and other regulated formats — from source data. Includes validation, reconciliation, and audit trail. Particularly relevant for credit bureaus and regulated lenders with high-volume reporting obligations.
CI/CD and DevOps
Automated build, test, and deployment pipelines. Environment promotion, structured logging, and rollback capability — so changes move from development to production safely and traceably.
Relevant engagements
- Automated report generation pipeline for Malaysia's leading credit bureau — XML and PDF at scale across multiple party types (InnovéLab engagement)
- IT transformation programme at CTOS Digital — cloud migration, data lake architecture, and platform modernisation (as GM Digital Factory)
- Enterprise architecture advisory for a major Malaysian manufacturer — covering WMS, ERP, Field Service, Retail POS, and CRM systems
What we build with
Container-native and cloud-agnostic. The stack is chosen for the problem, not for vendor preference.
Cloud & Infrastructure
AWS (primary) — including native data services (Glue, S3, Athena, Lake Formation). Container-native architecture (Docker, Kubernetes). Infrastructure-as-code. Deployable on any cloud or on-premise environment.
Data Engineering & Analytics
Cloud-native and open-source data platforms. Large-scale processing (PySpark, Databricks), analytical query engines (DuckDB, Trino), columnar storage (Parquet). Medallion architecture for tiered data lakes.
Application & Integration
Java, Python, REST APIs. Structured logging and audit trail by default.
Observability
Full-stack monitoring — system health, data quality, and business KPIs in a single view. Grafana-based dashboards, distributed tracing, log aggregation.
Security & Identity
SSO and role-based access control (Keycloak). WAF. Secrets management.
CI/CD
Automated build, test, and deployment pipelines. GitLab CI/CD or Jenkins. Environment promotion and rollback.
AI-Augmented Delivery
AI is embedded in our delivery process — solution design, software development, testing, and documentation. Opadeez includes a live AI builder that lets users configure workflows and data models in plain language.
Not sure which area fits your situation?
Most engagements start with a conversation about a specific problem, not a service selection. Tell us what you're dealing with and we'll tell you honestly whether we can help.
Start the conversation