What is Myrrix today began as the open-source project Taste in 2005. Providing implementations of basic recommender algorithms, it became a popular open-source recommender system because it offered a complete enterprise component.
Taste merged into Apache Mahout™ in 2008. The recommender implementations were revamped as "Taste 2" and expanded to include separate, distributed implementations based on Apache Hadoop™. These are among the topics covered in the book Mahout in Action in 2011.
Myrrix combines the popular real-time
Recommender API from Taste/Mahout, and integrates it completely it with distributed-style implementations begun in Mahout. It is "Taste 3", a continuing evolution of these recommender systems, from the primary author of these components, redesigned into a complete big-data-ready system.
Sean Owen: I am an engineer by trade, most recently at Google. I'm primary author of the Taste / Apache Mahout recommender engine, Mahout committer and PMC member, and co-author of Mahout in Action. Based in London, I have worked in early stage venture capital here. I have enjoyed building and talking about real-world recommender engines for seven years. It's exciting to bring into being a new project, and company, that will bring even better practical machine learning into the world.
There is a lot of history to make here -- UK-based co-founders are needed to move the product to 1.0, refine the algorithm implementations, and later, expand into other areas of machine learning.