After working at Google for four years after college, I’ve now been at Plaid for a bit under a year. People often ask - how does it compare?
What I miss
The developer tools
Google’s internal developer tools are really nice - here’s a few things I find myself missing or dealing with worse versions on a day to day basis:
- Flume (Apache Beam, sort of): The internal tooling around building large scale batch compute pipelines is top notch, and I didn’t appreciate it nearly enough at the time!
- Critique: The internal code review site. Nicely integrated with other tools and the turn-taking philosophy and CL Monitor chrome extension were top notch.
- BUILD files and Blaze aka Bazel: I encounter more mysteriously broken builds or build processes in a few weeks now than I did in years at Google. (A couple of my friends missed it enough to start a company, BuildBuddy, around it)
- Codesearch: Sourcegraph just doesn’t cut it - I think the tighter
integration really benefits Codesearch (“Find References” basically always
works, etc). And command-line searching with
ackisn’t even in the same category of tool.
- Cider and CitC: A really nice browser-based editor and source code that magically synced.
Google has one of the rare 401k plans that allows for the “mega backdoor
allow you to contribute after-tax dollars up to the $57,000 limit and convert
them to a Roth 401k. With the 50% match on the normal set of contributions,
you’d end up with
$19,500 * 150% = $29,250 in traditional pre-tax
$57,000 - $29,250 = $27,750 in converted Roth contributions
per calendar year! It’s a really great way to put money away quickly and it’s
feasible on high L3 (new grad) or L4 (one level up) Bay Area engineer
Memegen is Google’s internal meme site and cultural icon; at its best it’s a wonderful, hilarious place. (At its worst, not so much… read on)
What’s good out in the “real world”
The above is actually not that long of a list.
I’ve enjoyed working at a place that’s generally a hungrier for success. This doesn’t mean working crazy hours, but it does mean making the ones you do work count, and at this point in my career I’m not ready to rest-and-vest. (I actually literally did that at Google for three months after my last team got shut down, and hated it - but that’s a story for another time.)
Working on trying to make money
This is maybe a weird comment to make about an extraordinarily profitable company- but large parts of the Google engineering workforce are pretty far removed from the actual business of making money, leading to some classic issues like favoring new projects that’ll get you promoted. It’s refreshing to get back to focusing on business success instead of fuzzier metrics.
Learning transferable tech
I’ve been pretty happy to expand my skillset to use a bunch of different AWS tools (S3, EC2, Kinesis, Aurora, Elasticsearch) and others (Kubernetes) that’ll be relevant at other companies; Google’s stack was good but I sometimes got the feeling that I was over-specializing.
Memegen, especially post-Damore manifesto, became less of a fun way to poke fun at fancy lunches and instead became a battleground. Ultimately it resulted in more employee organizing and advocacy for important issues, which was good, but mentally it made it tougher to justify staying. (Rightly so! But my early 2010s halcyon vision of Google had been broken.)
Turns out I get to work with a lot of great people at Plaid too!