Racing through 2016 with so much on the front burner and yet it is timely to pause for a quick update on the launch of my new machine learning company, RaRe Technologies.
The Start of Something Exciting
I’ve heard from a few people who were confused when they received a recent newsletter from “RaRe Technologies”, when they signed up for a newsletter from Radim. Well, they’re very much related: RaRe = Radim Rehurek (in the same way gensim = generate similarities; admittedly, not the highest point of my creative efforts).
While some of our consulting projects at RaRe do involve Gensim (topic modeling, text embeddings…), many don’t. RaRe uses the full spectrum of data science tools to solve the hardest problems, both from the PyData ecosystem (keras, sklearn, pandas, luigi, matplotlib, seaborn, flask…) and non-Python world (mostly JVM; Elasticsearch, Spark, Hadoop, GATE, Mallet; VW…). As you know each world has its advantages and disadvantages. For Python, it’s usually memory issues once moving beyond toy-sized problems (hence streaming in gensim!); for non-Python, it’s cumbersome ops and a general lack of flexibility.
Gensim is and will remain open-source. Nothing changes. Well, one thing changed, the Gensim repository has moved from my personal github to under our company github. The Gensim community is growing quite nicely! There was an exciting gensim coding sprint at PyCon 2016 which was fun, and we met a lot of new gensim users.
Visit RaRe and let me know your thoughts on the new venture, or better yet, contact us for help on your next data science project.