Evaluation - Making Data Mistakes

Make it work, make it right, make it fast

I first heard this phrase from Gael Varoquax (one of the core scikit-learn committers) in a talk[1], and I loved the super-concentrated wisdom of it: Make it work. First of all, get it to compile, get it to run, make sure it spits out roughly the right kind of

Collaboration

How to reproduce your analysis

But first, an exciting announcement! I'm currently hiring for two Machine Learning Engineering roles, both doing interesting work with high-energy startups led by great people on good missions. Scroll to the bottom of this article for more details! How to reproduce your analysis “But didn’t you say you got

Data Science

How to Cheat at Data Science (with help from Centaurs and the Wizard of Oz)

What does it take to run a high-performing data science team? In this 15-minute video, I talk about two tools that I’ve found invaluable for doing just that: Wizard-of-ozzing is a technique for product discovery that can be very helpful for evaluating potential problems to work on [6:00]

Data Science

How to know if your recommendations algorithm is actually doing a good job

I led the team that built Channel 4’s recommender system for All 4 in 2016. It started out as a straightforward project. But after getting lost in a rabbit hole trying to devise a score for ‘provocativeness’ and ‘serendipity’, I learned the single most important lesson about data science.

Data Science

How to build a model that performs at chance

I’m going to tell you about a horrible screwup. Not my first and not my last, but a happy one, because although it was excruciating it was also instructive.