Leaving AMXL after 5 years⌗

What I learned from ~5 years at Amazon all of which were spent at an organization called AMXL. I’m moving to a new organization within Amazon, and I felt like it was a good time to reflect.

Reflecting on people⌗

Being negative and short with people is almost always detrimental to both you and them. If this is occurring figure out why and fix it asap.

Most requests will be labeled as “urgent”, very few will actually be urgent.

One of the best things a manager can do is help allocate your time well. There will be meetings you never sit in that have an outsized impact on what is worked on and how things are measured, especially long term.

General Reflections⌗

Good documentation is gold. Building in systems to keep it updated over time is gold x2.

Be cautious of your time… 5 years went very quickly, and I don’t see the train slowing down. You also have no idea where the end of the tracks are, enjoy it.

Clarifying questions are powerful & necessary, make sure as much context/information is made available before leaping into action.

It’s very important to understand inputs/outputs. People like to drop you into the middle of a complex problem/workflow which quickly becomes an unsolvable context trap if you don’t understand the start and the end. Also be careful of the XY problem.

Don’t allow my self worth to be entirely defined by work, I need other outlets.

A good metaphor is underrated. The ability to produce a good one quickly is even more underrated.

You should be failing 1-2x per year. This does not need to be out right crash and burn, but it should be something like “yeah we hit the limit of that system, looking back we should have done X, but now we NEED to do X”. This is how you learn, there is no book for this.

Data Reflections⌗

Under the right stressors, people will treat a data orchestration framework as cron.

People prefer a wide table rather than creating a data model. It’s important to be careful about this, See post on wide tables

Always assume that people have no understanding of statistics. Always assume your chart/graphic/UI is too complicated. This is NOT a knock on people. You are inherently waist deep in a problem that most others are only weakly aware of, if you can’t boil it down to something graspable that they can care about, that is your own fault.