As a Sr. Engineer, it doesn’t matter what you understand

It matters what you can explain

Brian Olson
4 min readMay 15

--

Disclaimer: All opinions are my own

Photo by fauxels from Pexels

I’ve worked with at least one “Wizard in the Corner” at every job I’ve had.

You know who I’m talking about — the most senior, smartest, most experienced person on the team. They can answer every question, they can solve every problem, and they’re assigned to every project. You can barely get time with them because every 15-minute slot of their day is filled with meetings with other people wanting time with them to solve their problems.

For whatever reason these individuals inevitably have desks in a corner of the office. I’m not sure why, but they do.

It’s great to have these folks around, and their experience can be invaluable. But you have to ask yourself — why does your team include engineers that aren’t them?

No, seriously — why? They can solve every problem, answer every question, and they’re involved in every project. Why do you need anyone else?

The obvious answer is scaling. A team of one can only accomplish so much. They can only have so many text files open, writing code as fast as you can think it. At some point, you find a limit to what that one critical person can accomplish.

This brings me back to the title — what that Wizard in the Corner knows isn’t actually all that valuable. Because access to that knowledge is rate limited by it being held by a single individual. What’s valuable from a Wizard in the Corner is how well they can explain things.

There are two kinds of Wizard in the Corner engineers: Redis clusters, and DNS servers.

Redis Cluster Engineers

Redis cluster engineers are the kind that spits back facts to you. You ask them, “My build is broken, what do I do?” And they reply, “Type cat magic_fix_.txt | grep "REMOVE" > input.txt". And you do that, and it’s fixed, but you have no idea where that magic fix came from, why it has REMOVE in it, and how the build gets input.txt.

Then production is down, and they tell you that you have to reboot the server at 10.10.10.2, and only later do you find out that it’s the login server that…

--

--

Brian Olson

Engineer, formerly at Amazon, currently at Google. All opinions are my own. Consider supporting here: https://devblabs.medium.com/membership