The National Geospatial-Intelligence Agency uses satellite images to learn all it can about every square inch of the Earth’s land masses, but one area of the globe that’s remained elusive for the agency is Silicon Valley.
That’s where Ben Tuttle, NGA geographer and director for the agency’s Silicon Valley outpost, comes in. One of his priorities is bridging the gap between the West Coast and Washington.
“Particularly with the outpost in the Valley, we’re looking at getting tighter with various industry professionals out there,” Tuttle told Federal News Radio during an April 13 USGIF workshop in Springfield, Virginia. “So the idea that they might be able to come and work with us to solve specific problems for a year, two years, three years, without having to make a whole 30-year government career out of it, and without necessarily having to move to D.C. and live inside the Beltway and get a top-secret clearance to do so.”
Tuttle said NGA is good at working with early startups and traditional defense acquisition processes that take place on a grander scale, but the people in that middle ground find it hard to work with NGA.
“So we’re trying to reach out and figure out how to bring more of that in, how to leverage more that’s going on in the commercial world and then bring results from that into our spaces as well,” said Tuttle, who is also director of development and operations for GEOINT (Geospatial Intelligence) services.
That’s something NGA Director of Plans and Programs Anthony Vinci also touched on during the workshop’s keynote speech. The workshop was held on the NGA’s Northern Virginia campus.
Vinci said NGA would be reaching out to find “non-traditional partners,” because ultimately what the agency has to do is “expand the scope of what we want to do.”
Tuttle said the way he’s connecting NGA with Silicon Valley is by bringing out public-sector people to translate government semantics and acronyms for people without a background in intelligence or defense.
“Just trying to make it easier, trying to put it in their lexicon, trying to be there to answer questions, so we’re not some weird, spooky black box of people in a dark building, but that there’s actually faces on the ground that can sit there and go ‘no, so what we meant was …,’ or ‘let me answer this question for you,'” Tuttle said. “Being able to go up at some of the events that happen out there, whether it’s at the universities or in the tech circles, and just put a face and an understanding on it that doesn’t make it so daunting to approach.”
Listen to Meredith Somers on Federal Drive with Tom Temin
NGA wants to shed light on its shadowy reputation, but industry can take some steps to help themselves in this process, agency analysts said.
Dan DeGennaro, a Synthetic Aperture Radar (SAR) imagery scientist, said during his workshop panel that one quality analysts and scientists looking at data are interested in is how much time is saved by using one particular service.
“When a service is provided, it is important to have thorough documentation of all the metadata, the values, and the attributes of those values,” DeGennaro said. “I believe one of our primary goals should be looking to reduce the learning curve as much as possible, by focusing on making the data ready to go almost right out of the box.”
Michaela Mesquite, senior GEOINT officer, echoed DeGennaro’s comments, saying during their panel that documentation should be easily accessible, and companies should be honest about their algorithm’s accuracy.
“We have a much higher tolerance for false positives, but a very low tolerance for false negatives,” Mesquite said, adding that it also means partners might need to be willing to “let us see that secret sauce.”
Tuttle said he doubted anyone would be open to revealing the ingredients of their secret sauce, however, that’s something NGA can work with.
“One of the things that our team does out there is sort of take the ‘OK, we heard what you said from our folks here, and we can get at your intent,’ but with something that we understand the community out there is able to swallow,” he said. “So rather than show me your secret sauce, I would say be ready to show me the actual statistics of how effective your algorithm is, which a lot of them haven’t done in some cases, and now they’re going ‘oh OK, so I need to actually do statistical studies of … false positives, false negative things like that.”
Don’t bring another PowerPoint to a meeting, Tuttle said, and be ready to let analysts sit down with your system and run through a workflow to see what comes out the other end.
“It’s not showing you the secret sauce, but it’s letting them actually see that it works and they got a result that makes sense to them,” he said.