Facebook-Acquired Startup Utilizing Computer Vision To Scale & Automate Mapping

Google Earth Inspiration

Fresh off watching The Billion-Dollar Code (highly recommended), a German TV miniseries on Netflix based on actual events on how two entrepreneurs went to court with their company Terravision against Google over infringement rights about the Google Earth algorithm, I was inspired to find a startup or company using artificial intelligence (AI), machine learning (ML) and/or deep learning (DL) for geographical mapping to write about. Luckily, the search was not a long one and I came up with a worthy candidate quickly.

Founded in 2013 by Jan Erik Solem, Johan Gyllenspetz, Peter Neubauer, and Yubin Kuang but acquired by Facebook in 2020, Mapillary is the street-level imagery platform that scales and automates mapping using collaboration, cameras, and computer vision, making images and map data available to visualize the world and help develop cities, maps, and automotive.

Before Facebook’s acquisition, signs were good anyway for the Malmö, Sweden-based startup as it had raised a total of $24.6 million in funding over four rounds with capital sourced from the likes of Playfair Capital and LDV Capital.

The magic is in the Mapillary platform, which allows end-users to host, process and publish street-level imagery and map data, is it can provide solutions for:

• Keeping maps detailed and up to date

• Speeding and scaling up mapping

• Making imagery available as open data

Judging by what Mapillary was offering, it is little wonder Facebook was interested, as it gave Zuck and co. a chance to ramp up its imagery and mapping data and add more scope to compete with Google.

“From day one of Mapillary, we have been committed to building a global street-level imagery platform that allows everyone to get the imagery and data they need to make better maps. With tens of thousands of contributors to our platform and with maps being improved with Mapillary data every single day, we’re now taking the next big step on that journey.”

 — Mapillary Co-founder & CEO Jan Erik Solem, 2020

Had it not been for Mapillary’s talented founding team, Facebook may have never been interested. Without delving into this counterfactual too much, though, let’s take a brief look at the quartet.

Mapillary’s Founders

Now Engineering Director of Map Building at Facebook, former CEO of Mapillary Jan Erik Solem is an entrepreneur and computer vision expert with an interest in large-scale computer vision problems.

Having founded Polar Rose (acquired by Apple in 2010), he then worked for Apple as an Engineering Manager and Computer Vision Researcher and also spent nearly a decade as an Associate Professor at the Mathematical Imaging Group at the Department of Mathematics, Lund Institute of Technology/Lund University doing research in applied mathematics and computer vision.

Solem has a Ph.D. in Applied Mathematics from Lund University.

Also employed as an Engineering Manager at Facebook is Johan Gyllenspetz. Starting his career as a Software Engineer at Swedish security solutions company Axis Communications, like Solem, Gyllenspetz has been the co-founder of several tech startups.

Gyllenspetz obtained a Master’s in Computer Science from Lund University.

Peter Neubauer is also an Engineering Manager at Facebook. The co-founder of Neo4j and a number of popular open-source projects such as Neo4j, OPS4J and Qi4j, Neubauer loves to connect things, write novel prototypes and throw together new ideas and projects around big data and society-scale innovation.

Another serial founder of startups, he has a BS in Wood and Paper Technology at the KTH Royal Institute of Technology.

An Engineering Manager at Facebook as well, the fourth co-founder is Yubin Kuang, a man passionate about researching and deploying large-scale computer vision and ML solutions to mapping.

Prior to Mapillary and Facebook, Kuang had two short stints at Apple as a Software Intern. He has a Ph.D. in Applied Mathematics from Lund University, where his work on geometric problems in computer vision, image retrieval, semantic image segmentation, optimization, structure from the sound was an asset for his future career.

It’s evident that the four can now focus on what they’re good at as engineering managers at one of the biggest companies in the world, employing computer vision for street-level imagery, rather than doing some of the other, less exciting, distractions that are endemic when running a startup. And with Facebook’s backing, Solem, Gyllenspetz, Neubauer, and Kuang can help one Silicon Valley giant make gains with another Silicon Valley giant when it comes to mapping every area of the globe.