Danti uses natural language processing to make searching Earth data simple
Commercial and defense organizations have an Earth data overload problem: they collect vast amounts of heavily siloed data, with few ways to quickly synthesize insights and deliver them to the people who most need it.
Danti, a startup emerging from stealth today with $2.75 million in funding, wants to solve this problem with a powerful natural language search engine for location-based data. This should enable non-experts to quickly sort through huge amounts of diverse data and find the answer they need.
The search engine could completely change how location-based data is used and understood. On its website, Danti provides a number of possible queries an end user could pose. These span the mundane queries of an insurance underwriter (“What are the risks for property at 2340 River Bend Rd”), to the more urgent questions of a warfighter (“battle damage assessment in Irpin in last 48 hours”).
The startup is headed by Jesse Kallman, whose career has spanned unmanned aerial systems, satellite data, and geospatial intelligence. He eventually landed at Airbus, where he led a new business unit focused on developing geospatial market software, before moving to autonomous aviation startup Xwing, where he was VP of commercialization and strategy.
“I’ve learned the same lessons over and over and over again,” he said. “It was never about the drone, it was never about the satellite. It was always about, what is the question that the user is trying to answer, and how do you get the right information [to them] based on what it is?”
The solution to the problem, he says, is in the search layer. Just as the first search engines revolutionized the ability for the non-expert to find useful information online, Danti could similarly unlock Earth data end users — so its founder hopes, anyway.
“That’s where we were really inspired and what we really want to try to accomplish,” Kallman said. “How can you build something just purely at the search layer, not going into […] the really nitty gritty geospatial work, for the [National Geospatial-Intelligence Agency] analyst or even better, the downstream deployed army user who’s just trying to answer a simple question about a property?”
To accelerate development of the search engine technology, Danti closed a $2.75 million pre-seed funding round led by Tech Square Ventures with participation from Radius Capital. The round also included Philip Krim and Raven One Ventures, SpaceVC, Overline, Tareyton Ventures, Jordan Noone, Keith Masback, and Jeff Crusey.
Danti’s team stands at less than 15 people and is scarcely four months old, but the startup is moving fast. In April, Danti won an open challenge from the National Security Innovation Network, which asked competitors to pitch solutions to reduce the time-intensive data search process for the National Geospatial-Intelligence Agency. Kallman said the follow on feedback from NGA, as well as commercial customers, through the open challenge and other beta testing has been “incredible.”
“That’s where we are at this stage. It’s really, really getting deep with many different types of users to really understand what’s valuable to them.”
Danti uses natural language processing to make searching Earth data simple by Aria Alamalhodaei originally published on TechCrunch