0xdata, the startup behind the H2O open-source machine-learning project, has secured a new funding round to the tune of $8.9 million. And as it looks to grow, the startup is ditching its original name and going with H2O instead.
Other open-source companies have made that change — look at Chef, MongoDB, and Redis Labs, for example. Now the team behind H2O are aligning with the name they’re more likely to be known for, having accumulated a user community for the project.
“The open-source movement has definitely helped us get phenomenal distribution even in last three, four quarters we’ve been selling the product,” SriSatish Ambati, chief executive and a cofounder of H2O, told VentureBeat in an interview.
Other open-source tools for machine learning and a more complex process called deep learning have attracted attention from developers. There’s Deeplearning4j, which Skymind pushes. There’s PredictionIO. There’s also Apache Mahout, which Hadoop distribution vendors Cloudera, Hortonworks, and MapR support.
And then there are startups with machine-learning services developers can incorporate into their apps with application programming interfaces (APIs); AlchemyAPI, Skytree, and Wise.io fall into that camp. Ambati likes the idea of having an API, and H2O would like to create one, but for now it can benefit from being available for free as an open-source project. H2O can then make money by providing support and maintenance.
What’s more, H2O integrates nicely into the workflows of data scientists who are familiar with the R programming language. H2O itself is partly written in R — as well as Java and Python — and people can run H2O functions from R once they’ve also installed the H2O package for R. And part of that is by design: Like others who have built on R, Ambati wasn’t satisfied the performance of R on its own.
“I was literally fuming and sending notes to the founders of R,” he said. He ended up collaborating with R core member John Chambers, among others, to develop H2O.
The project has grown popular since those days, and the startup by extension has picked up business. The startup now claims 20 production customers, including Cisco, eBay, Nielsen, and PayPal.
Ash Bhardwaj, Riverwood Capital’s Michael Marks, and Rakesh Mathur led the new round in the startup. Nexus Venture Partners, Transamerica Ventures, and Jishnu Bhattacharjee also participated. To date H2O has raised $13.6 million.
H2O started in 2011 and is based in Mountain View, Calif.
The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here