PredictionIO, a company aiming to become the “MySQL of prediction”, has just raised $2.5 million in seed funding to open source machine learning, and allow developers to build predictive applications with just a few lines of code. Investors in the seed round include Azure Capital, QuestVP, CrunchFund and Stanford.
As well as receiving funding from Stanford, PredictionIO are a part of the Stanford-affiliated StartX Accelerator Summer roster, receiving $100,000 of resources from the accelerator’s impressive portfolio of partners and guidance from over 200 affiliated serial entrepreneurs.
With the impressive funding round, PredictionIO are fuelling their grand ambition- to be the machine learning server behind every application. Although they’re not the first to offer to offer machine-learning-as-a-service (companies like Google Prediction and Skytree have that one covered), their unique selling point comes from the fact they’re open-sourced. Simon Chan, Prediciton IO’s CEO, told Dataconomy one of the principle reasons for open-sourcing their technology is to give users flexibility. “Enterprise companies with in-house data scientists can add unique features on PredictionIO, they can even build their own algorithms on our platform if they want,” he said. “This is something “black box” solutions cannot offer. And we find it extremely important for enterprise to build up in-house data competitive edge on open platform.”
The PredictionIO community already comprises over 4,000 developers who are contributing to PredictionIO’s open source development on Github, and hundreds of applications in industries such as e-commerce, retail, fashion recommenders and news media. Chan feels the community to set to expand rapidly; “The potential of building Machine Learning algorithms on open source PredictionIO is unlimited,” he remarked.
“One day, we’ll see “self-driving car” engine on PredictionIO. Developers can build self-driving intelligence into auto with just a few API calls.”
Their plans with the new funding are certainly ambitious. But Chan’s ambitions spread beyond the scope of his own venture and out to the whole machine learning community. “Today, all applications have a Database server behind. Soon, all of them will need a Machine Learning server, to personalize content, to predict user behavior, to become smarter basically.”
“Machine Learning will become as common as database.”
(Image credit: PredictionIO)