Project Overview
| Client Industry | Finance / Consumer Lending |
| Business Type | Regional consumer lending company processing personal and auto loans, ~150 internal users |
| Project Duration | 26 weeks |
| AI Service Provided | Product Modernization |
| Technologies Used | React, TypeScript, Node.js, .NET (legacy), PostgreSQL, SQL Server (legacy), AWS (ECS, RDS, S3), Terraform, Auth0 |
The Client Challenge
The client’s loan origination platform had been built 12 years earlier on a legacy .NET Web Forms architecture with a SQL Server backend, and had been patched and extended incrementally ever since by a rotating cast of contractors. It processed active loans and couldn’t simply be taken offline or rewritten overnight.
The compounding problems:
- The platform’s original architects were long gone, and current staff had incomplete knowledge of significant portions of the codebase, making even small changes risky and slow.
- New regulatory reporting requirements couldn’t be implemented in the existing architecture without significant rework, putting the company at risk of compliance gaps.
- The UI was built for desktop use only with no responsive design, while loan officers increasingly needed to work from tablets during in-person and remote client meetings.
- Page load times for loan application review had degraded to 8–10 seconds as the loan volume and historical data grew, since the original architecture wasn’t designed for the current data scale.
- Onboarding new engineers took months due to the codebase’s age, inconsistent patterns, and lack of documentation a serious risk given how few people understood the system.
- The company was evaluating acquisition offers, and technical due diligence flagged the legacy platform as a significant risk factor affecting valuation discussions.
A full rewrite was too risky given active loans in process and the operational disruption it would cause. The client needed genuine modernization without stopping the business.
Our Solution
Air Brite Labs executed a phased modernization using the strangler fig pattern: building a modern React and Node.js application alongside the legacy .NET system, migrating functionality module by module, and decommissioning legacy components only once their modern replacements were proven in production.
Key outcomes delivered through the modernization:
- Modern, responsive interface — loan officers can now review and process applications from tablets and laptops, not just desktop workstations, with a completely rebuilt UI.
- Improved application review performance — page load times for loan review dropped dramatically through both frontend rebuild and backend query optimization.
- New compliance reporting capability — built directly into the modern architecture, resolving the regulatory reporting gap that the legacy system couldn’t accommodate.
- Reduced institutional knowledge risk — the modern codebase follows current best practices with proper documentation, meaningfully reducing the “only two people understand this system” risk that had concerned leadership.
- Zero disruption to active loans — the phased migration approach meant loans already in process throughout the modernization continued without interruption or data loss.
- Improved technical due diligence position — the modernized platform materially strengthened the company’s position in ongoing acquisition discussions.
Technical Approach
Migration Pattern: Applied the strangler fig pattern a modern Node.js API layer was built in front of the legacy system, initially proxying requests to the existing .NET backend, then progressively replacing individual modules (loan intake, document management, underwriting review, reporting) with modern implementations one at a time.
Frontend Rebuild: Built a new React and TypeScript frontend to replace the legacy .NET Web Forms UI, module by module, allowing loan officers to transition to the modern interface incrementally rather than facing a disruptive all-at-once UI change.
Database Migration: Data was migrated from the legacy SQL Server database to PostgreSQL in phases, using a dual-write strategy during transition periods to keep both systems synchronized and allow safe rollback if issues arose with any migrated module.
Compliance Reporting: New regulatory reporting requirements were built directly into the modern Node.js backend and PostgreSQL data model, avoiding the architectural constraints that made this impossible in the legacy system.
Authentication: Migrated authentication to Auth0, replacing the legacy system’s custom authentication and enabling modern security practices (MFA, session management) the old system couldn’t easily support.
Infrastructure: Deployed the modern application layer on AWS ECS with infrastructure defined in Terraform, while the legacy system remained on its existing infrastructure until each module’s migration was complete and validated.
Risk Management: Every module migration included a defined rollback plan, and legacy components were only decommissioned after their modern replacements had run successfully in production for a defined validation period critical given the platform’s role in processing active, real-money loans.
Implementation Process
- Discovery & Requirement Analysis (Weeks 1–5): Conducted an in-depth audit of the legacy codebase, working with remaining staff who had partial system knowledge to document business logic that existed only in code, not documentation a substantial undertaking given the system’s age.
- Prototype / PoC (Weeks 6–8): Built the strangler fig proxy layer and migrated the lowest-risk module (reporting dashboards) first to validate the migration pattern before touching core loan processing functionality.
- Development (Weeks 9–19): Migrated loan intake, document management, and underwriting review modules sequentially, rebuilding each in the modern stack while keeping the legacy system as a fallback until validation was complete.
- Integration (Weeks 20–21): Migrated authentication to Auth0 and completed the compliance reporting module using the now-modernized data model.
- Testing (Weeks 22–23): Ran parallel validation for each migrated module, comparing outputs between legacy and modern systems on real (anonymized) loan data to confirm functional equivalence before full cutover.
- Deployment (Weeks 24–25): Completed the final module migrations and fully decommissioned the legacy .NET application, with the SQL Server database retained temporarily as a read-only historical archive.
- Optimization (Week 26 and ongoing): Tuned database query performance and refined the new UI based on loan officer feedback collected throughout the phased rollout.
Key Features Delivered
- Fully rebuilt, responsive loan origination interface usable on tablets and laptops
- Modern Node.js and PostgreSQL backend replacing the 12-year-old .NET/SQL Server stack
- New regulatory compliance reporting capability built into the modern architecture
- Auth0-based authentication with MFA support replacing legacy custom authentication
- Zero-disruption phased migration preserving continuity for active loans throughout
- Significantly improved loan application review performance
- Modern, documented codebase reducing institutional knowledge risk
Business Results
- Loan application review page load time dropped from 8–10 seconds to under 1.5 seconds
- Zero active loans were disrupted or delayed throughout the 26-week modernization
- New regulatory reporting requirements were successfully implemented, closing a compliance gap the legacy system couldn’t address
- Loan officers gained tablet and remote access capability for the first time, supporting more flexible in-person and remote client meetings
- New engineer onboarding time dropped substantially, based on internal team feedback, due to the modern, documented codebase replacing undocumented legacy logic
- The modernized platform was cited as a positive factor in subsequent technical due diligence during the company’s acquisition discussions, removing a previously flagged risk
Technology Stack
| Layer | Technology |
|---|---|
| Frontend | React, TypeScript |
| Backend (Modern) | Node.js |
| Backend (Legacy, decommissioned) | .NET Web Forms |
| Database (Modern) | PostgreSQL |
| Database (Legacy, archived) | SQL Server |
| Authentication | Auth0 |
| Infrastructure as Code | Terraform |
| Cloud Infrastructure | AWS ECS, RDS, S3 |
Why the Solution Worked
The modernization succeeded because it respected the real constraint at the center of the project: a live lending platform processing active loans couldn’t tolerate a risky, disruptive rewrite. The strangler fig approach building the new system alongside the old and migrating incrementally meant the business kept running normally throughout, with a validated rollback path at every stage.
Investing real time upfront to extract undocumented business logic from the legacy codebase, working directly with the few remaining staff who understood parts of it, was equally critical. Skipping that step would have risked silently losing important logic during migration a serious problem in a regulated lending environment where correctness isn’t optional.
Future Scalability
The modernized platform is built to support the company’s continued evolution, including:
- Adding API-based integrations with credit bureaus and underwriting data providers, now straightforward with the modern Node.js architecture versus the legacy system’s rigid structure
- Extending the compliance reporting module to accommodate future regulatory changes without the architectural constraints of the old system
- Building a customer-facing loan application portal that integrates directly with the now-modern backend
- Supporting continued technical due diligence needs as the company moves through its acquisition process, with a platform that no longer represents a flagged risk
Because the legacy system was fully decommissioned rather than left in place alongside the modern platform, the company now has one clean, current architecture to build on rather than two systems to maintain going forward.
Final Outcome
A 12-year-old, undocumented legacy platform that had become a genuine business risk was modernized into a fast, compliant, and maintainable system without a single disruption to the active loans it processes throughout the transition. What was flagged as a liability in acquisition due diligence is now a platform the company can confidently build on, regardless of how those discussions conclude.


