I joined Wayfair as a full-time employee on August 1st, 2011, to run internal search and customer recommendations, after a brief stint as a consultant. I have other responsibilities now as well, but over at the company tech blog, http://engineering.wayfair.com, I've written mostly about the search and recommendations systems. We've done a lot over there in a year. Let's recap a few highlights:
- We wanted to deploy a Python web service, so we figured out a few things about web operations with Python, which we expressed in both comic-book and serious/detailed form.
- We set up a KISS-method recommender system based on simple correlation metrics, as a proof of concept.
- That worked, so we started down the road of more involved recommendations, at first with Markov clusters.
- Jeff Bertolucci of Information Week picked up that piece, did an interview with me, and wrote an article about it: http://www.informationweek.com/big-data/news/big-data-analytics/240007850/online-retailer-uses-dna-research-to-connect-with-customers. The crack team of data scientists is delighted to see their work highlighted in this way!
- We improved internal search in various ways, including hacking Solr and contributing our patch back to the open source community.
- We wrote some classifiers and other machine-learning programs for various purposes, and then linked one of them up to the search capability.
It's hard work, I suppose, but mostly it just feels exhilarating and fun.