Workday

Snacks, pay. My old org was great and felt like one big team.

No remote work, and yet can't get anything done in the office with all the chit-chat.

They seem to view infrastructure and build systems as a complete afterthought, and had between 1 and 2 engineers maintaining a build system used by 80+ teams.

Leadership seems more focused on pumping out worthless prototypes to get kudos from their manager than actually building the tools the organization needs to be successful.

The entire duct-tape and band-aid operation is held together by a tiny minority of very talented and hardworking engineers, without whom no team in the company would be able to release software (except the random outliers who never migrated off of some even older system, or acquisitions doing their own thing).

Now they're cracking down hard on anyone who might potentially be enjoying their job or building anything worthwhile, by moving towards a fully data-driven culture. It's a perfect recipe for justifying very stupid decisions.

For example, should we automate our deployments? Anyone with the slightest experience, and even anyone who can manage to use Google, will say "yes, obviously." But that's not data-driven!

First, we must do everything the wrong way to gather data. Then, when we eventually get around to doing it the right way, we'll have our measurements and can use data to make a decision.

Except now you're maintaining a bunch of manually created prototype services that some brilliant person decided to let users onto (early feedback!).

The users are now reliant on these systems, and the work to build long-term solutions becomes further and further away. It's a speck on the horizon until it disappears entirely.

Now the org is asking why everything is taking so long. Users are complaining that basic functionality such as "being able to log in" isn't working. Workday is paying hundreds of thousands of dollars a year to have skilled engineers manually resetting passwords.

Do we really need to run an experiment to know that we should follow general quality standards? Does no one understand that if you're 20% into a project and users are using 20% of the functionality, something is horribly, horribly wrong?

Maybe this stuff is not obvious to everyone, or maybe they just shut up and come in.

But believe it or not, there are standard engineering practices that should be taken for granted.

• We shouldn't need data to understand automated tests are valuable.
• We shouldn't need data to know automated processes are more efficient than having engineers clicking buttons.

This isn't the half of it. I enjoyed most of my time there, though, and it's highly dependent on what team you're on and your manager.

I'd also stay away from "high investment" areas like AI-related work. They seem to be full of ladder climbers who will do things that hurt the company for their own gain.

There is politics, although I just write code and ignore it. I can't imagine anything more stupid than people who should all be on Team Workday actively working to harm each other.

But I will say this: I've heard from new hires that it's actually pretty good and competent compared to what else is out there. I had only worked with exceptional teams before this, so I don't know if it's a good baseline. It damn well makes me worry about how bad it can get outside.

When you continually put strong pressure on the next level down, the pressure on engineers becomes extreme. We were continually told to ignore security requirements when dealing with customer data because the pressure was so intense.

Also, you guys need to learn how to hire good engineers and then do that. Right now, there are like 3 bad engineers and 4 decent engineers for every 1 or 2 good engineers.

Also, fund the build system for God's sake. It's absolutely insane that every team I've been on has had more people than does the team that maintains a system used by 80 teams and without which half the company would grind to a halt.

If you were actually smart, you'd work on stuff like making developers more efficient. Every team is wasting time writing nearly the same garbage tooling to run in the CI. Does that seem efficient? I hope that's a rhetorical question because my firsthand experience is that it's a gigantic waste of time.

Probably a full half of everything employees are doing is a waste of time. But since they won't address common-sense issues without the data, it just becomes a continuous cycle of 'wait until it's an emergency, then maybe we'll do something about it.'