Résumé Skews for A/B Testing

#

In a previous post, I described my system for dynamically generating a résumé skew from an org-mode document made in literate programming style to emit a LaTeX file, which in turn is used to create a PDF.

To be honest, while I thought it was (very fun) overkill at the time, it has already been really nice to be able to easily update a thing or two in the org doc, push it up to SourceHut, and then run this blog's pipeline to get the updated version on the site. Today, though, after some feedback that my résumé was incredibly verbose, a wall of text, and an overwhelming amount of content (jk, it was very good and kind feedback!), I decided to work up a new skew for it, using bullets for my work history instead of paragraphs.

I am calling these two bullets and narrative for the moment. You can see the narrative form here and the bulleted form here. The process of doing this highlighted the actual practical utility of this new system: I was able to keep all of the narrative content and all of the bulleted content in the same source file, just tagged with different skews. It is trivially easy to generate one or the other, and in fact the default Nix build spits them both out.

All of this makes the prospect of A/B testing the two formats feasible! Obviously it won't be a perfect comparison, since no two roles are alike, but AS THERE'S NO POSSIBLE WAY YOU MIGHT HAVE GUESSED, I maintain a list of potential roles to apply to in an org-mode document. Adding a quick property to note which were applied to with the narrative skew vs the bullet is ezpz, as is tracking which roles get rejections with no further interview requests. Obviously the sample size will be somewhat limited, but it still could result in some interesting data, which I will eventually post here!