October 19, 1998

SciComp licenses software to Dutch bank

---

Marla Dial Austin Business Journal Staff

An Austin company that makes "software that writes software" has inked a licensing agreement with a far-flung European bank.

SciComp Inc., a 4-year-old startup, is releasing its first product: SciFinance, a tool that takes the programming work out of pricing a complicated type of investment called derivative securities. A large Dutch bank, MeesPierson, has licensed the product for use in its Amsterdam headquarters.

Financial terms of the deal were not disclosed.

Derivatives are a type of security investment based on the value of other underlying securities. Bank employees can spend days or weeks writing software code for highly complex derivative pricing models, but SciFinance is designed to do the same work up to 10 times faster.

MeesPierson employs about 4,000 people and operates in 25 countries. The bank serves a range of business clients: entrepreneurs, managers and treasurers of mid-size and large international organizations that trade in commodities markets.

"They're a pretty sophisticated user of derivatives," SciComp founder and President Elaine Kant says.

"We can learn a lot by working with them," adds Curt Randall, vice president of technology applications. "We expect good enhancements to the software through working with them."

SciComp's technology is designed to help customers produce their own complex, mathematical software, without manual computer programming.

SciFinance "is software that writes software," Kant says. "Using our system, they develop specific kinds of software that they can distribute throughout the company."

Although SciComp has a few other customers, MeesPierson is the first bank to announce its relationship with the company publicly. Kant says other banks -- including several large investment organizations that tested SciComp during an 18-month period -- don't want their names disclosed, for competitive reasons.

The MeesPierson licensing is another milestone for SciComp, which recently secured $1.5 million in first-round venture funding from a New Jersey firm, Verticality Investment Group LLC, to bring its product to market.

The startup previously received $2 million from the U.S. Department of Commerce Advanced Technology Program, says Kant, who holds a math degree from the Massachusetts Institute of Technology and a doctorate in computer science from Stanford University.

Kant, whose interests in using computers to automate mathematical processes led her to start the company, didn't originally plan to target the financial industry. But Randall -- a physicist Kant met while both were working for Schlumberger -- had several contacts on Wall Street, where scientists in the early 1990s were providing derivative pricing models to help traders predict stock market movements.

Randall saw an opportunity for SciComp to apply technology in the same way.

Kant says SciFinance is one of three ways banks can do derivatives pricing. The others are employing someone to write pricing code by hand or buy "library packages" that price only certain kinds of derivatives deals.

SciFinance, however, creates a language that can address a variety of derivative scenarios.

"Because it's software that writes other software, the banks can use that to price deals we've never even thought about before," Randall says.

The basic licensing fee for SciFinance is $100,000 per user per year.

Here's how the software works: a quantitative analyst tells the computer about a particular finance problem in about half a page of text -- including specifications, equations, boundary conditions and final values for the derivative.

SciFinance uses that explanation to generate thousands of lines of code and spits out a model for solving the problem.

Stathis Tompaidis, assistant director of the Center for Computational Science at the University of Texas, is using SciFinance in his own research and graduate business classes.

"Without necessarily knowing how to program themselves, [students] are able to develop software that can price complex instruments," Tompaidis says.

For banks -- which often write their own, proprietary software for new pricing instruments -- SciFinance can slash the development cycle by one-third to one-half, he estimates.

Randall says banks that use the product can examine its code themselves to understand the pricing models it yields. That means they don't have to rely on a mysterious mechanism when huge amounts of money are at risk.

What's more, SciFinance can be used to help banks assess and hedge the risks they incur after selling derivatives.

(c) 1998, Austin Business Journal