Skip to content
All work

FinTech|2022 – 2024

MeridianLink Consumer

One application from any channel, decided by rules and booked once.

MeridianLink

Consumer

The consumer loan origination system behind every channel: apply, decide, book.

A loan officer walking a member through an application at a branch desk
Product imagery: MeridianLink, from meridianlink.com.
Through
Bhavna Software
Role
Mid Level Developer

Overview

MeridianLink Consumer is a cloud-based loan origination system for banks and credit unions: the system of record that takes a consumer loan application from any channel, decisions it, and hands the approved loan on. It covers personal and auto loans, HELOCs, credit cards, small business and indirect loans.

I contributed to Consumer from August 2022 to September 2024 as a software engineer at Bhavna Software India, on workflows, decisioning rules, partner integrations and the features that bring one applicant's channels together. The highlighted parts of the map are where that work sat.

  1. Where applications start

    Applicants, members, dealers and brokers

    • Online (this case study)
    • Mobile (this case study)
    • Branch (this case study)
    • Indirect (this case study)auto dealers
    • Brokers and TPOs

  2. Point of sale

    The application front end

    • MeridianLink Portalsucceeded by Access, Oct 2023
    • Mortgage Accessmortgage point of sale

  3. Origination

    Systems of record for loans and accounts

    • MeridianLink Consumer (this case study)consumer loan LOS
    • MeridianLink Openingdeposit account opening
    • MeridianLink Mortgagemortgage LOS, formerly LendingQB
    • PriceMyLoanpricing and eligibility engine

  4. Around it

    Hundreds of partners through the MeridianLink Marketplace

    • Core banking
    • Credit checks
    • Identity and fraud
    • eSignature
    • Document management
This case studyConsumer is the system of record for consumer loans from every channel.

The platform in numbers

16M+
auto loan applications
8M+
unsecured loan applications
5M+
new accounts opened
1M+
home loan applications

Lifetime figures published on meridianlink.com, read October 2026. They describe the platform, not my work.

Impact

Four things that changed for lenders and underwriters. Each row names the problem, what I did about it, and which layers of the stack the work went through.

  1. Dynamic

    configurable workflows

    Workflows lenders can shape

    Lending teams at banks and credit unions

    The problem
    A lender's process varies by product and institution, and a fixed workflow forces staff to work around it.
    What I did
    Developed dynamic workflows in the origination system so steps follow the product and the lender's own configuration.
    The loan lifecycle
    • UI (yes)
    • API (yes)
    • Data (yes)

  2. Automated

    decisioning rules

    Decisions made by rules, not by hand

    Underwriters

    The problem
    Applications that a rule could decide still waited in a queue for a person.
    What I did
    Developed automated decisioning rules, so qualified applications move on and underwriters spend their time on the ones that need judgement.
    Decisioning
    • UI (no)
    • API (yes)
    • Data (yes)

  3. 4

    channels: mobile, online, branch, indirect

    One application, every channel

    Applicants and lenders across four channels

    The problem
    The same applicant could arrive through different channels, and the rules had to behave the same in each.
    What I did
    Built application-consolidation features across mobile, online, branch and indirect channels, so one record and one set of rules apply.
    Constraints
    • UI (yes)
    • API (yes)
    • Data (yes)

  4. Integrated

    third-party services

    Partners plugged into the flow

    Lenders using third-party services

    The problem
    Credit, identity and other partner services have to fit the origination flow without breaking it.
    What I did
    Built third-party integrations into Consumer's workflow, inside a platform whose marketplace connects hundreds of partners.
    • UI (no)
    • API (yes)
    • Data (yes)

Four of the four went through more than one layer of the stack. The layers come from what my resume names for this work: React and TypeScript screens, ASP.NET Core and C# APIs, and EF Core over MSSQL.

The product

A consumer loan's path through the origination system, from application to booking.

  1. Apply

    Online, mobile, in branch or through a dealer

  2. Consolidate

    One application record whatever the channel

  3. Credit and data

    Credit checks and partner data pulled in

  4. Decision rules

    Configurable rules and logic trees

  5. Approve or review

    Qualified applicants advance; the rest go to underwriting

  6. Cross-sell

    Pre-qualified products offered in the same session

  7. Book

    Loan handed to core banking through integrations

What it is
A cloud-based consumer loan origination system for banks and credit unions, covering personal and auto loans, HELOCs, credit cards, small business and indirect loans.
Decisioning
An advanced decisioning engine with configurable rules, visual logic trees and real-time data. Automated approvals move qualified applicants on at once and leave deeper review for the loans that need it.
Channels
A mobile-first digital application, in-branch applications, and indirect lending. MeridianLink added indirect lending capability when it acquired Teledata Communications (DecisionLender) in November 2020.
Compliance and reporting
Embedded controls and audit trails, fraud and identity integrations, and built-in reporting with Insight Lite and up to three years of origination data.
Lender console · Application searchRedrawn · fictional data
Lender console application search with filters for branch, status, dates, loan officer, funding and messages, above a results table

Swipe to see the whole screen.

Lender console · Application searchThe search staff use to find applications: branch, account purpose, status, created and funding ranges, loan officer, ACH verification and messages, with results they can assign. Redrawn for this page field for field from MeridianLink's published screenshot of its console; the data is fictional.

Screens

Two screens of the lender console, drawn for this page with fictional data: the application queue and a decision on one application. Pick a callout to see which part of the screen it explains.

Screen 1 of 2

Application queue

Application queue: Every channel in one queue, one record per applicant.

Every channel in one queue, one record per applicant.

The lender's view of incoming consumer loan applications: filtered by channel, decided by rules where possible, consolidated when the same person applies twice, and assigned to an officer when a person is needed.

Role

  • Software engineer at Bhavna Software India, on MeridianLink's products from August 2022 to September 2024.
  • Created core components, and coded across HTML, CSS, JavaScript (ES6 and later), TypeScript, React, C#, VB.NET and ASP.NET Core.
  • Wrote maintainable, testable code with NUnit and Moq.
  • Worked with the client on requirement analysis for stories and tasks, system design, and product enhancements.
  • Managed deployments and monitoring through Azure Monitor and Azure Blob Storage.

Constraints

  • Consumer lending runs across mobile, online, branch and indirect channels, and decisioning rules had to behave the same in all of them.
  • Lending is regulated: changes had to keep the embedded controls and audit trails intact.
  • Partner services for credit, identity and fraud sit inside the flow, so an integration could never block an application.
  • Legacy VB.NET and MSSQL code shared the system with newer ASP.NET Core and React work.

Architecture

Applications from every channel reach the same origination services. An ASP.NET Core and VB.NET service layer runs the workflow and decisioning rules over MSSQL, with partner integrations for credit and identity, and the approved loan moving on to core banking.

Every channelOnline, mobile, branch, indirectOrigination servicesWorkflow & decisioning rulesASP.NET Core / VB.NETService layerMSSQLApplication & loan recordsCredit checksIdentity & fraudCore banking

How I worked

Straight from the client
Requirement analysis for each story and task happened with MeridianLink directly, so system design and enhancement ideas came from the people who own the product.
Sprint rhythm
Feature planning, sprint discussions and peer code reviews, with work tracked in Jira and designs handed over in Figma.
Tested as it ships
Unit tests in NUnit with Moq for the services behind each change.
AI assistants, used deliberately
GitHub Copilot and ChatGPT to speed up feature work, troubleshoot issues and support code reviews.

Results

More consistent underwriting and compliance handling with faster processing across the consumer lending pipeline: workflows that follow the lender's configuration, rules that decide the routine applications, and one application record whatever the channel.

4
channels on one application record
2
years on the MeridianLink stack
1,900+
institutions on MeridianLink by 2021

Stack

  • ASP.NET Core
  • C#
  • VB.NET
  • Web API
  • EF Core
  • MSSQL
  • React
  • TypeScript
  • JavaScript (ES6+)
  • NUnit + Moq
  • Azure Blob Storage
  • Azure Monitor
  • Jira
  • Figma

Next project

MeridianLink Mortgage

Read the case study

MeridianLink

MeridianLink Mortgage

ASP.NET Core · C# · VB.NET · Web API