Paul Bass Photo
Daniel Abadi’s new Science Park company snagged $9.5 million to start hiring people fast — in Cambridge.
Abadi (pictured) and a team of fellow Yale researchers developed a hot new software program called “Hadapt.” It has the potential to revolutionize the way companies store and analyze huge amounts of data.
Investors thought enough of the program’s potential to offer millions of dollars in start-up capital to launch Abadi’s company, Hadapt Inc., to further develop and sell it. A team of investors recently gave the company $9.5 million.
Now the company is scrambling to hire 30 software developers and marketing people. At its new office in Cambridge’s Kendall Square. Only a handful of employees will remain in New Haven.
All the potential investors made it clear that they would participate only if the company located in either Cambridge or Palo Alto.
“That question came up very early,” Abadi, a Yale assistant professor of computer science, said in an interview in his spare Prospect Street academic office. “The investors said they had no interest in investing if we stayed in New Haven.”
As New Haven pursues an “eds and meds” job-creation strategy, the story of Hadapt’s rise follows the script for policymakers hoping to grow new companies by commercializing university research. But it reflects how in some cases the city may not yet have reached the point of being able to translate that script into actual new jobs.
The hopes begin with the brains and insights of entrepreneurial academics like 31-year-old Abadi.
Jumping Through Hadoops
Abadi, who now lives downtown, came to Yale four years ago. He earned his PhD at MIT studying “scalable data management.” That means figuring out how organizations using thousands of computers at once can store, analyze, and use huge amounts of data.
Abadi offered a hypothetical example of what kinds of questions that field explores: How does Walmart, say, test the wisdom of moving basketballs from a top shelf to a lower one, and moving footballs up to eye level? It feeds copious records of sales transactions from different seasons, some from regular stores, some from test sites with lowered basketballs, into banks of computers, which then seek to crunch the numbers to see if the tests produced better sales.
It’s a lot more complicated than it sounds.
Abadi planned to continue that work as a faculty member at Yale. With fellow professor Avi Silberschatz, four PhD students, and a posse of undergraduates, he started a lab known as “DR@Y,” for “Database Research at Yale.”
Abadi wanted to make an impact on his field. So rather than invent software completely from scratch, he decided to work with a popular open-source program called Hadoop. A developer named Doug Cutting started creating Hadoop based on a 2004 paper Google published about its own approach for getting thousands of machines to work in parallel to allocate and analyze data. Google and Walmart keep their codes secret. Cutting’s made his public — or “open source” — so techies around the world helped him continually develop it. Hadoop took off after Yahoo hired Cutting and a team of developers to improve it — and keep it public. Today “even old stodgy firms like Sears,” Bank of America, and GE use Hadoop, Abadi said.
In his DR@Y lab, Abadi’s team worked on a new software program based on Hadoop that could enable companies to process and analyze all that data much faster. Unlike Hadoop, which uses raw “unstructured” data, the new program would draw from data already arranged in tables. Larger companies like Verizon have such tables.
Abadi and company called the program “Hadapt.” (As in: adapt Hadoop.) They tested it. The results: It worked some 50 times faster than regular Hadoop.
So they now had a proprietary program to bring to market. If they could find the money.
In keeping with Yale’s mission of putting tech research to commercialize use, Abadi brings an entrepreneurial streak to his research.
“I want my research to be used by eBay, Fidelity,” he said. “It’s way more fun to have your impact while you’re still alive. Especially when you’re up for tenure.”
The opportunity came almost by accident. A venture capitalist had come across Abadi’s research results. He alerted Yale’s Office of Cooperative Research (OCR) that it may have a golden prospect on its hands. A School of Management student named Justin Borgman was working in that office at the time. He saw the notice. He contacted Abadi, asked for meeting along with people from OCR. Let’s start a company, Borgman said.
So in the fall of 2010 they started Hadapt. They set up shop in Science Park. Borgman serves as CEO. Abadi and some of the lab members have equity stakes and work part of the time for the company, part of the time at DR@Y.
It turned out “dozens” of potential investors made inquiries about financing Hadapt’s start-up. A team led by NorWest Venture Partners and Bessemer Venture Partners prevailed, ponying up $9.5 million to get the company off the ground.
That was in October. Now Hadapt is hustling to fill 30 positions. The company has some beta clients already and hopes to get version 1.0 of its software program to market some time in January. It’s keeping a handful of people in a room at Science Park, but otherwise hiring and working in Cambridge.
“Here’s the basic issue,” Abadi said. “Database software is a very, very specialized industry. There have only been one or two dozen database companies that have ever done anything successful.” Maybe 200 people at most worldwide are qualified for Hadapt’s most specialized developer positions, he said. And they tend to be located in Silicon Valley, near Stanford; in Seattle, near Microsoft and the University of Washington; or near MIT in Cambridge. Hence investors’ insistence on leaving New Haven.
On The Map
Is that bad news for New Haven?
Disappointing, perhaps, but understandable. With a definite upside.
So says Anne Haynes, who runs New Haven’s Economic Development Corporation (EDC).
“We should be proud of Hadapt’s success,” she said. “People will know it started here. As much as it’s sad to lose a company like that, they’re going to have great success and always be known as a Yale and New Haven-minted company.”
So while New Haven can’t compete to provide workers for this one highly specialized field right now, it will get more of a reputation as the birthplace of hot tech companies, she argued. That can help attract new venture capital for other tech businesses, gradually growing the sector.
EDC, meanwhile has been organizing “orientation activities” for venture capital firms, leading tours and mailing information about Yale’s and the city’s growing tech sector.
Abadi said he believes New Haven can become “a center for technology innovation,” including database management. “There are a lot of smart people around New Haven.”
He, for one, hopes to stay around New Haven. In addition to hitting a venture capital home run, he just had another life-changing event: He’s engaged to be married.