Excel Add-In
Migration
Inherited a legacy VB.NET Excel add-in firing raw ODBC queries. Migrated it to scalable Spring Boot APIs, introduced HazelCast distributed caching, replaced a paid MSI generator with an open-source WiX pipeline, and rebranded the add-in — all without touching the core financial logic.
The Problem
HighRadius had a mission-critical Excel add-in used by enterprise finance teams to upload and fetch consolidation reports. The add-in was written in legacy VB.NET and communicated with database servers using raw ODBC queries. This direct connection created huge security risks, high latency, and database locks during peak financial periods.
Additionally, generating the installer (.MSI file) for client machines relied on a commercial, closed-source generator with expensive annual license seats. The system was fragile, slow to update, and couldn't easily be integrated into a modern CI/CD pipeline.
The Solution
I migrated the data retrieval architecture from raw database connections to a secured, microservice-based layer built on Spring Boot. Direct queries were replaced with highly optimized REST APIs that use the Strategy pattern to handle varying payloads dynamically based on report structures.
To handle heavy multi-dimensional financial queries without overloading the core database, I integrated HazelCast distributed caching. This cached complex reporting trees, reducing database load and speeding up fetching actions.
To solve the installer problem, I completely replaced the paid packaging software with an automated build script using the open-source WiX Toolset. This pipeline compiles and packages the add-in during the build, producing a signed MSI installer automatically. I also restructured the UI to support the company's new branding, giving the add-in a modern look.
Tech Stack
Details
| Company | HighRadius |
| Role | SDE-2 |
| Type | Legacy Migration |
| Domain | Fintech · Architecture |
Key Outcome
Eliminated recurring installer licensing costs by migrating to an automated WiX pipeline. Achieved up to 5x faster data fetching for end-users using HazelCast caching and stabilized the database under heavy concurrent loads.