part time engineering position at Whatify.com, early stage tech/data start-up in New Haven
mission: democratize randomized experiments (aka a/b testing)
beach head market: ~30m entrepreneurs on aggregate ecommerce marketplaces (etsy, eBay, amazon) who want to increase traffic
team: Jason Abaluck, SOM economics professor. Paul Nichols, seasoned start-up technologist Cara Mae Cirignano, FES enviro econ grad.
how it works: we a/b testing elements of shops and listings for search engine optimization. We use machine learning & natural language processing / image modeling to first automate test design, and then to pool data from all tests to deliver robust results. Example: you sell orange blankets on etsy. You sign up on whatify (thats all that’s required as a user). We a/b test a few keywords, and then install 100 optimized words across your shop, resulting in a permanent bump in traffic.
revenue model: monthly subscription for ongoing testing & optimization
what’s happening at the moment: we’ve run a/b tests for ~100 etsy sellers to date, opening up the waiting list for the next batch in late November
web dev tech: Ruby on Rails on Heruko , although considering a switch to another language
what we’re looking for: student technologists who have the following:
- programming experience in machine learning and/or web development
- a deep desire to develop a real product getting put in front of users asap
- a very bare minimum of 5 hours a week to devote to this
- a strong commitment to completing the project
- interest in one of the three projects outlined in webpage linked below, or related project (we can tailor)
what we can offer:
- compensation
- mentorship
- a full view inside the guts of an early stage start-up
if you’re interested
fill out the form below - we’ll get back to you within a few days
and/or send questions to caramae@whatify.com
Project Descriptions
Please note:
- you don’t need deep experience in similar projects
- we can tailor to your interests & skills
Machine learning - text or images
Jason's model draws on vectors derived from word2vec, which has been trained on Etsy listings, and outputs lists of the top tags for our user's listings. You'll improve the model by identifying new word corpora on which to train word2vec, use the resulting vectors in Jason’s model, and compare the new output to what we already have to identify where & how improvements have been made.
Web development with the Etsy API / Ruby on Rails
You’ll create a webpage where users can easily review per-listing tag recommendations & reject tags they don’t like. Design skills not necessary
Web development with another API or language
You'll build an MVP version of this web app, identical in functionality to the one we’ve already built with the Etsy API, but with the Amazon or eBay API. MVP functionality includes the ability to run a simple a/b test on eBay listing titles.