Analyze This

A Boca Raton tech company seeks an antidote for anthrax and dotcom failure

No one sits at the reception desk in Suite 325 of the Mizner Office Complex in Boca Raton. The waiting area for Sengent, whose company logo of eight cubes creates a three-dimensional grid, has become something of a game room where stressed-out computer programmers can unwind and walk away for a moment from the ceaseless glow of monitors.

That's because, instead of a secretary waiting in the reception area, a dartboard hangs invitingly. Dozens of tiny holes ring the board, suggesting that too much time in front of a computer might be bad for aim or worse for eyesight.

Company President Gioel Molinari, a tall thin man with thick brown hair and sideburns, smiles as he runs his hand over the pockmarks in the wall. "If only everyone here weren't blind," he says jokingly. Dressed in jeans and a white button-up shirt with the Sengent logo on it, Molinari talks with a slow deliberate eloquence, as if he rehearses the lines in his head right before speaking. The Fort Lauderdale resident is a self-taught computer programmer with the kind of entrepreneurial spirit that thrived when websites that sold dog food traded on NASDAQ.

He wears designer eyeglasses and a week's growth of facial hair, pulling off that New Economy-changes-everything shtick so perfectly, you'd expect him to have been a Business 2.0 cover boy in the late '90. But he was a bit too young back then.

Now at the tender age of 28, Molinari has successfully launched a South Florida tech company during an economic drought that turned dot-com into profanity. Earlier this year, he raised $425,000 from four accredited investors in a seemingly dry market.

That, however, is the dull part of Molinari's story. This is a guy, after all, who helped to develop software that may one day identify an antidote for anthrax, smallpox, Ebola, and SARS.

A Carnegie Mellon University graduate with a degree in chemical engineering, Molinari came to South Florida in March 1999 to take a computer programming job at Xcelerate, a Fort Lauderdale company that provides e-business consulting. As the bottom fell out of the Internet economy, Xcelerate and other tech companies languished.

But Molinari saw opportunity. He and his fellow coworkers had an interest in working with one of the computer industry's emerging technologies, distributed computing, also known as grid computing. The technology allows huge computational problems to be divided among hundreds or thousands of computers. This network of machines, known as a grid, can provide the computational power of a supercomputer.

While grids have for years been used by academics as a way to increase power without escalating costs, the technology has only just begun to find its way into commercial markets. That's where Molinari's Sengent -- whose name comes from combining sentient agent -- comes into play.

After individual stock traders lost fortunes on Wall Street, thanks to biased brokerage advice and willy-nilly trading, Molinari saw a future for grid computing in the investment industry. Risk analysis and investment forecasting, both of which require intense levels of computation, can take a single computer hours, even days, to calculate. Spanning that workload over a network of machines would decrease calculation times, Molinari speculated, giving investment banks and hedge funds an edge over the competition.

But Sengent needed startup capital. Yeah, you've read this story before, the one about the group of guys, all under 40, out to grab their piece of the Internet fortune. Where this tale of high-tech enterprise takes a twist is where Molinari received a call from his uncle, Jonathan Rothberg, president of a genomics research company and founder of the Rothberg Institute for Childhood Diseases. The nonprofit institute in Guilford, Connecticut, researches possible cures for Tuberous Sclerosis Complex, or TSC, a genetic disorder affecting children that leads to benign tumors in the eyes, brain, heart, kidneys, lungs, and other organs.

Rothberg wanted to find a way for his institute to perform computerized drug screening without having to invest millions of dollars in hardware. His nephew had a solution: a volunteer grid. In exchange for startup capital, Molinari agreed to a challenge from his uncle. His company would create a drug-screening process to identify antidotes for potential weaponized diseases, including anthrax, smallpox, and Ebola.

Computerized drug screening, a maturing technology that identifies chemical compounds for future lab testing, is an ideal project for grid computing because it requires vast amounts of computation. A library of 100,000 chemical compounds can be narrowed down to as few as 12,000 possible candidates for drug testing. "Basically, we play video games," explains Bonnie Gould Rothberg, the institute's medical director and Rothberg's wife. "We simulate to see whether any of these drug molecules can fit like a key into a lock. What we're trying to do is get the computer to simulate chemical interactions."

In October 2001, Jonathan Rothberg gave Sengent 30 days to develop the software. Molinari and his ten employees worked day and night. "It was madness," Molinari remembers. The project became known as Drug Design and Optimization Lab, or D2OL, and employs software architecture originally intended to crunch financial data. It launched on November 15, 2001.

What makes D2OL different from Sengent's financial software is volunteerism. Unlike large corporations with networks of computers on hand to crunch data, the nonprofit project must rely on the computer power of average Internet users. A PC owner interested in participating downloads a small program that runs computations when his computer is idle, dedicating otherwise wasted resources to the project. The program runs computations, for example, when the screen saver is activated, making it unlikely that the user would realize D2OL is running in the background. When the program finishes its computations, it automatically uploads the results and receives a new assignment.

1
 
2
 
All
 
Next Page »
 
My Voice Nation Help
0 comments
Sort: Newest | Oldest
 
Loading...