The Jonah Group

• Overview
• Clients
• Case Studies
• By Industry
• By Service

Credit Risk Modeling - Credit Vantage Network (CVN)

The Problem

Through a partnership between Algorithmics (a member of the Fitch Group) and Citigroup (the largest financial institution in America), the objective of Credit Vantage Network (CVN) is to provide access to a global suite of single-obligor credit models and default risk analysis tools.

While working with another vendor, a series of delays and overbudget iterations prompted Algorithmics to halt development on the CVN project. At that point, the system was still incomplete and unstable for production. The Jonah Group was asked to take over the entire development and was given the mandate to complete the remaining features with production quality within two months for CVN's first release. A major sales effort was also simultaneously underway to promote the product, targeting major financial and lending institutions worldwide. Further setbacks would be difficult to tolerate in this context.

Our Responsibilities

• Project planning and timeline estimation
• Management of a team of 4 people
• Participation in architecture and design reviews for the pending features
• Refactoring existing design and code to enhance software quality
• Addition of extensive unit test suite
• Assist integration with the wider business systems and activities
• Ongoing Communication with our client's Lead Architect, Project Manager, Business Leads and 2 Software Developers

Back To Top

The Solution

Jonah provided a project manager, a senior technical architect and 2 software developers to complete the CVN product with the following features, components and technologies:

Algorithmics was responsible for quality assurance, acceptance testing, and promotion to production

Features

• Calculation of the Probability of Default (PD) from a comprehensive suite of credit models
• Web-based user interface to visualize the financial health and risk of lending portfolios
• Monthly update of all public traded companies' financial information and global stock markets (a feature that was upgraded to daily update)

Back To Top

Components

• Web application to display and analyze Probability of Default and financial portfolios
• Probability of Default Web Services to perform PD calculations and portfolio analysis
• Standalone Windows application to retrieve global market data and public traded companies' financial information from Fitch registered sites

Back To Top

Technologies

• Apache Struts
• Web Service and SOAP
• MS Excel
• Formula One and JExcelAPI (Java libraries to read, write, and modify Excel spreadsheets)
• XML / XML Schema
• Java / Servlet / JSP
• Apache Tomcat Server
• Oracle 10g

Back To Top

Unique Challenges, Solved

• The original system had large amounts of defects and it could support only low volume of transactions and concurrent users. Jonah refactored areas of the original system design that required high throughput and concurrency, and imposed a rigorous test harness to ensure the system's correctness.
• Faced with the changing needs of clients to whom Algorithmics was pitching the product, the project team was required to adapt rapidly and seamlessly to changes in scope. Short development cycles were introduced to provide flexibility on project planning, marketable features, and usability fixes. By establishing vertical development processes, features for potential buyers were released without sacrificing quality.
• In addition, Jonah designed and built the Probability of Default (PD) Web Service, which enabled Algorithmics to expand their product offering to their clients, providing another revenue channel.
• With the project previously over budget, years in development, and the client’s management without much insight into the completion of major project features, there was mounting pressure to have a version of the product ready for sales and public use. Jonah planned semi-weekly integration build cycles along with two to three month go-live iterations to demonstrate success on a regular basis. Increased responsiveness to client questions and concerns (relative to the former vendor) and a frequently updated project extranet - containing the weekly stable version of the software - provided the required visibility and confidence that the project was on the right track.

Back To Top