Ryan Clough, A Professional

The talk @daikeshi an I did at #kubecon2019 is now up on youtube! If you want to see why & how we are providing @kubeflow as a centralized resource for our ML practitioners at @SpotifyEng, check it out! youtube.com/watch?v=m9Xhsn…

wow, even more depressing to see our stock skyrocket over this

Please reach out if your team has opportunities for my colleagues and friends, you won't be disappointed in their work and passion 🖤

Massive layoffs at Spotify today 😞 While I was not layed off, I have many colleagues whom I would gladly work with again that will be looking for jobs. MLEs, EMs, PMs all with full scale ml or ml infra experience, if your team/company is looking for it.

the google cloud logs viewer is the worst. I just spent 10 minutes trying to figure out how to just copy a multi line stack trace. I gave up and just sent a link. It has so many bells and whistles when 90% of the time I'd be better served from being able to just cat a log

PS if anyone wants to help me get on bsky, HMU :)

returning briefly to give yall another ML Ops meme

Not to mention the hilarity of "salient lines of code".... Outside of really low level programming, great code isn't like a programming interview, it involves interactions across multiple systems, impactful abstractions, etc that cross numerous files....

I could not imagine the humiliation of having to do this after another 1.2k of my coworkers just left and I'd signed up (or was coerced due to immigration issues) to be "hardcore" for Elon

I want to dig in for a bit on this set of Slack replies from a Twitter manager talking about undoing some of the layoffs.

@hanneshapke @casassaez Certainly other engineers will continue to contribute at other companies but not as likely that they'll have the sort of environmental mandate a company the size and culture of Twitter has

@hanneshapke @casassaez Lots of open source projects are historically comprised of the biggest companies that make use of them.... But for production ML, Twitter is def among one of the bigger ones, definitely a big brain drain for open source caused by the lay offs

Was cleaning up my calendar and got this sad response from one of my meetings, RIP to much of Twitter's ML Platform team 💔

@casassaez I can only imagine 😔 sorry you have to deal with such chaos and the loss of so many of your coworkers

@fnord2vec 🥲 my email look a lot like this all week long

@bernhardsson I forgot backstage is open source now.... I have no idea if the OSS version of this has it but it would be in here if it did backstage.io/docs/features/…

@bernhardsson Our docs system in backstage does this at Spotify (def launched after your time), you can select text and there's some kind of widget that will open a GH issue on the repo where the docs are committed, filed out with the highlighted text. Not sure what powers it...

Can someone build a widget where if you select any text in developer documentation it pops up something like "would you like to leave feedback on this?" and lets me submit something?

The role is really fun because you get a wide view of all the different kinds of ML being done across Spotify (at least 30+ teams with ML projects!) and help shape he direction of the ML platform. We also do a lot of lighter consultations with more mature teams.

If this sounds like something you'd be interested in doing, and have experience with implementing best practices for production ML pipelines, especially with Tensorflow and TFX, let me know (and ask questions), because we're now hiring a second "me" 😁 lifeatspotify.com/jobs/senior-so…

This past year at @SpotifyEng I've started working as a solutions engineer within our ML Platform - Every quarter I work directly with 1 or 2 different teams to help them achieve their production ML goals by helping hem onboard, implement and sometimes augment our platform tools.

Why fight with infrastructure when you can fight with an abstracted DSL instead?

The dumbest thing I hear is “Facebook listens to your conversations” My coworkers were a bunch of risk-averse MBAs who’d never build something that cool. So why do you see ads for something you talked about? It’s because your life is predictable & the ad algorithm is that good.