Thanks for visiting Building Bayesian. We hope you like it and it expands your brain and all that good stuff.

Who is ‘Us’, here?

This blog is co-written by two St. Louis Data Scientists: Shane Caldwell and Tyler Kirby. We’re best buds who spend a lot of time at work together building machine learning systems. It’s sort of a Neil Gaiman/Terry Pratchet situation, but instead of writing about the end of the world and the absurdity of planning in a chaotic universe, we’re interested primarily in machine learning with a bayesian slant and the absurdity of planning in a chaotic universe.

If you find yourself in desperate need of determining who wrote what, understand that Shane is the one who tries to be funny and Tyler understands the English language and knows when you say who vs whom vs whomst.

Why are you writing about Bayesian ML?

Our hope is after reading enough of these articles you will understand that a bayesian framework is not just intellectually interesting or academic, but extremely practical in machine learning environments. If you read this blog and apply this framework to your own work, there is no greater compliment. The systems we build in ML come with inherent uncertainty and risk, so being able to formalize those notions and be rigorous in their analysis is indespensible to building stuff that actually works.

Plus it makes you sound really smart, and you can’t put a price tag on that.

Who should read this?

Ideally everyone interested in Bayesian ML, but hopefully there’s a lot here for anyone working with large, messy datasets trying to rigorously make predictions with them.

One day someone might ask you questions about your level of confidence in the fancy new system you developed. Or the fancy systems level of confidence in its predictions! Or someone might just ask you to A/B test a website and you want to knock it out of the park and get that attaboy you’ve been looking for.

This should help with that.

Will you write about other stuff?

We’re not completely married to Bayesian topics, and reserve the right to talk about other stuff, but idelogically we’re pretty Bayesian and that will likely inform how we write about other topics. So maybe. Idk.

This website is powered by fastpages 1.

  1. a blogging platform that natively supports Jupyter notebooks in addition to other formats. ↩