PredictionIO has closed $2.5 million in seed funding. Investors include QuestVP, Azure Capital, CrunchFund, Stanford – StartX Fund, Kima Ventures, IronFire, Sood Ventures and XG Ventures. PredictionIO is an open source machine learning servicer for programmers and developers.
PredictionIO (http://prediction.io) has quietly raised $2.5M seed funding to “open source” Machine Learning for developers to build smarter stuff. Investors include QuestVP , Azure Capital (investor of KISSmetrics, VMWare etc), CrunchFund, Stanford – StartX Fund, Kima Ventures, IronFire, Sood Ventures and XG Ventures etc.
Accelerators: We are part of Mozilla WebFWD, 500Startups and StartX.
WHAT IS PREDICTIONIO
– “MySQL of Prediction”
– An Open Source Machine Learning server for programmers/developers to build smarter stuff with just a few lines of code.
WHO ARE USING PREDICTIONIO
– Powering hundreds of applications – ranging from individual developers to international enterprises, in verticals such as e-commerce, fashion suggestion, retails, food delivery, video portal, news media, app stores etc.
– A community of 4000+ developers engaged in the open source product development (https://github.com/PredictionIO/PredictionIO)
Some User Showcases:
Le Tote – http://www.letote.com
Le Tote, a clothing subscription/rental service, is using PredictionIO to predict customers’ fashion preferences.
PerkHub – http://www.perkhub.com
PerkHub, an enterprise SaaS company that powers perks and group buying programs for the world’s leading companies, is using PredictionIO to personalize product recommendation in their weekly emails.
THE PROBLEM WE ARE SOLVING
Building Machine Learning into products is expensive and time-consuming. Only companies like Google, Amazon, LinkedIn, Twitter can afford a huge team of PhDs/data scientists. Nowadays, with tools like R, Hadoop, Spark, Mahout, Scikit-learn, companies need to build their own Machine Learning deployment infrastructure and server.
The problem? If it doesn’t make sense for most companies to build a database server in-house, then why should they need to build a Machine Learning server from scratch?
WHY IS IT A BIG DEAL?
PredictionIO aims to be the Machine Learning server behind every application. (remember MySQL in the early day of database revolution?) Building Machine Learning in software will be as common as search soon with PredictionIO.
COMPETITORS AND OUR UNIQUENESS
PredictionIO is differentiated by being open source compared to closed “black box” MLaaS services or software (Google Prediction API, Wise.io, BigML, Skytree) and developer-friendly compared to “research projects” with steep learning curves build by data scientists for data scientists (Vicarious, Numenta, Mahout).
– no developer like black-box solution (that’s why Hadoop, Docker are so popular
– give away free Machine Learning server to destroy, and then re-define, the Machine Learning market (think MySQL in the database market)
– Design for production deployment – a college student tweeted that he built a personalized app discovery engine using PredictionIO within 30 minutes
– No Machine Learning knowledge required
HOW DOES PREDICTIONIO WORK?
Steps for developers:
1. Download and Install PredictionIO or launch PredictionIO CLOUD on AWS Marketplace
2. Stream app event data into PredictionIO through REST APIs or with a few lines of code using SDKs(similar to Kissmetrics, Google Analytics)
3. Create prediction engines using PredictionIO UI
4. Retrieve prediction results through REST API calls
WHO ARE WE?
Simon Chan – Co-Founder and CEO
– Founded 3 startups in the past
– Lived in Hong Kong, Guangzhou, London and SF Bay Area in the past 10 years
– PhD candidate of University College London (UCL)
– B.S.E. Computer Science from University of Michigan, Ann Arbor
Donald Szeto – CTO (Stanford, UCBerkeley); Thomas Stone – VP of Enterprise (UCL-PhD); Kenneth Chan – Founding Engineer (UCBerkeley), Justin Yip – Data Scientist (Brown-PhD)