These days we have a ton of geospatial data coming from the growing number of satellites orbiting our atmosphere, but it takes some serious processing power and engineering skill to turn that data into something useful.
Some former Uber engineers, Sina Kashuk and Isaac Brodsky, who helped build the mapping system at Uber, decided to put their knowledge to work on the problem, and what they came up with is a fast serverless product that takes advantage of some tools open source tools they’ve created to take that data from the source and into applications where it can be put to work.
Today, the startup emerged from stealth with the Melted platform and $1 million in pre-seed funding.
Fused is a three part system. It takes data, which is fed from satellites, sits in storage warehouses, and runs it through its platform to make it usable. The ultimate goal is to create visual representations of things like weather, deforestation, or crop data, and put the data to work in applications people use like Excel, Airtable, or Notion. What separates it from its predecessor is the simplicity and speed of processing.
Kashuk, who is Fused’s co-founder and CEO, says the company took some time to build the current solution. In fact, after leaving Uber, they started another company in 2019 called Unfolded.ai that was looking at a similar data visualization problem, but it was very quickly acquired by Foursquare.
“The main challenge we faced with Unfolded was that it was very difficult to get to market because the majority of the platform was open source and people said ‘why should I pay you when I can just use the open source,'” said Kashuk at TechCrunch.
This was a business model problem, but they also faced limitations because they were trying to run a front-end solution locally on a laptop, and as the data scaled, it became too slow. As they were thinking about their next venture, Kashuk and Brodsky saw this opportunity with data coming from the growing number of commercial satellites and realized that back-end data processing remained a challenge and could be a business.
In particular, the founders saw the maturation of serverless computing, where the vendor can handle all the back end infrastructure, as an opportunity to help customers process that data and use it faster and more efficiently. “So the serverless movement combined with massive amounts of data coming from the growing commercial satellite industry led us to believe that the industry was ready for this kind of change,” Kashuk said.
The platform is essentially a middleware processing layer that helps transform geospatial data into something more consumable. It consists of several open source pieces and a serverless processing engine. It’s the last one where the company makes money. Every time data reaches the API gateway in the serverless backend, Fused gets paid.
Fused workbench in action. Image Credits: Melted
The entire system is built with Python, which is widely used by data scientists and developers. The system starts with a kind of visualization template called a user-defined function, or UDF. This part is open source that allows the community to create these patterns, such as crop yields, and then using other pieces of the Fused platform, they can begin to analyze the data, such as the amount of wheat produced by the state of USA or any geographic region you would like.
They can then transfer this data to other programs for further analysis or create data visualizations based on the data. One of the big differentiators here is the speed at which Fused creates these visualizations or processes changes to them. Instead of taking hours, it takes seconds, according to the company.
One of the key platform pieces involved in creating visualizations is the Fused Workbench, which is available starting today. Workbench, while partially open, only works in conjunction with the back-end API, he says, but it lets you interact with the data to see different aspects and see your changes almost instantly.
The company launched last year and has been working with early beta customers before going live today. The $1 million in pre-seed funding came from Fontinalis Partners along with various industry angels.
