Using statistical inference to let you find jobs in design, engineering, sales, and more: autotagging is now live (Week 60)
Here’s a brand new feature: autotagging jobs with words like Design and HR using text analysis and statistical inference 💥
So if you want to see only the design jobs at startups I endorse, well here they are:
See how UX Researcher is tagged Design even though the word “design” isn’t used? That’s statistical inference for you. Only about 1 in 6 jobs provide function or department metadata. But out of the 8,000+ jobs that we have in the database, that’s a bunch, and it’s enough to train our automatic classifier to make tag predictions for the rest.
At this point I also want to do a shout out to Brian Suda for early pointers to the best statistical techniques to use, right at the outset of this autotag journey. Insanely useful. THANK YOU!
All the usual Job Garden features work too: add boards to your watchlist (signed-in only), then filter by location and autotag. Watching boards from Veen and Downing and just want the hardware jobs on your list? We got ya.
Want a weekly email or RSS feed of hardware jobs only in LONDON? Yup. Check it:
I have wanted to be able to browse by job function forever because it makes JG about 10x more useful. But had to wait until the system had seen a bunch of jobs.
Turns out you can do interesting things with data once you’ve got enough of it, and we’re only scratching the surface.
A quick note about the new help pages:
I’ve always enjoyed writing for the web because it has a different structure and rhythm than writing, say, a chapter in a book or prose that is meant to be read out loud.
Recently Monzo published their Tone of Voice guide and it’s fantastic: comprehensive and practical all at once. I kept it open in a separate tab all the while I was writing.
It strikes me that there is room in the world for a Strunk & White of writing for digital: long form newsletters, UI copy, and everything in-between. Monzo’s guide would be a great place to start.