Four Months as an Analyst at Amazon

My perspective switching to the Supply Chain BI team.

A little Background

The first four months have been interesting. Last week, Amazon laid off 18,000 people. Fortunately, nobody on my team or, I think, my whole org got the axe. This really hit home the fact that jobs come and go, and that I should be prioritizing employment that allows me to work on what I’m interested in and gain meaningful skills.

I’ve been working ~60-70 hours a week, mostly because I want to be good (I want to be the best) at my job, I’m learning a lot, and I’m about to transition to a BIE (given that 10 different people and our VP signs off that I am capable/deserve it). I have a deep desire to fulfill every request of me as quickly and accurately as possible which is both a good/bad trait. I’m learning how to prioritize my work and focus on one thing rather than bouncing around.

While I would be lying if promotions or wide-spread recognition isn’t nice, I’ve found that individuals reaching out to tell me they appreciate my work and my attentiveness to detail means the most to me. I also derive a lot of meaning from the 1:1 conversations I have with the people I view as mentors.

-Personally, I’ve been enjoying spending time with friends (Even reaching out and
getting together with people I haven’t seen in years). I’ve also been actively
organizing and spending time with my family; this is something I didn’t do in
high school/college outside of major holidays, now in my early 20’s,
I’ve attempted to rectify this before it becomes a major regret in my 30s/40s+.

-I’ve been studying poker, I played for the first time at a casino recently:
my hands were shaking the entire time, and I lost ~$300 over the course of
3 hours. I got an espresso machine, and I’ve enjoyed using it
(I am not a coffee snob!).

-One of my child-hood friends got married. It was surreal remembering
playing pool basketball and watching Iron Man at my 4th grade
birthday party/sleepover, now he has a wife.

-I’ve been continuing to run my book club at work, this month we are reading
Talking to Strangers by Malcolm Gladwell.

-I’m working on being the best son, brother, friend, peer, and partner I can be.

What has been not so appealing/difficult?

  1. Creating reports, models, and queries for a new part of the supply chain that I’m less familiar with has been challenging. There are a variety of teams that create complex models that generate complex outputs to optimize different parts of our network. We consume these outputs into our Datawarehouse. I must interpret those results, figure out if they make sense, figure out if they make sense for our specific network, and then create visualizations that make those results easily accessible to PMs/Leadership.

  2. Supply Chain Program Managers and their bosses each have a unique set of jargon and metrics for their silo/role in the supply chain. Understanding what everything means and what it means in the context of other jargon is quite complex!

  3. We have close to 500 tables containing over a Petabyte of data on our Redshift Custer (Datawarehouse). Figuring out what tables are useful and accurate for my queries has taken months to develop. I now use a subset of 30-40 tables for almost all asks. 100/500 tables cover almost all asks from supply chain teams, the other 400 tables are used for the mist niche of requests and queried a handful of times per year. However, those tables still had to have long initial syncs and then subsequent data syncs (daily) that use up valuable/limited cluster resources.

  4. I don’t like when other people’s reports break or when the numbers “aren’t quite right” for a dashboard. This usually entails digging through a ton of SQL (Sometimes poorly written) or trekking through all the steps in someone’s data pipeline, or worst of all digging into their scrappy solution where we are web-scraping into a an excel file and reformatting with VBA. At the end of the day, this is an important part of the job, but I feel like its thankless/non-value add work.

  5. Dealing with issues that have been kicked down the road: Poor organization of our tableau dashboards, scrappy process’ that few people actually know about/can fix, POOR DOCUMENTATION (Please add comments in your queries, an info tab in your dashboards, and links to your ETL process), poor data validation/inadequate testing, and improper error handling/alert systems/code reviews.

What have I enjoyed?

  1. I’m quite passionate about writing complex and performant SQL and helping to educate others do the same. I’ve written a couple documents about optimizing queries to scan less data and run faster, how to use the explain plan, and reached out to folks to help them write better queries. I’m going to be the mentor for one of our new interns starting next week, so I’ll be learning what the best way to get these concepts across quickly is. I’ve also written some stored procedures for database admin tasks, identified and fixed tables with improper sort/dist keys and retention policies, and adjusted poor performing ETL jobs.

  2. I’ve gotten to work on some fascinating projects. How do we position/spread inventory as close to customer demand as possible? How do we improve our buying/forecasting systems and our responses to overbuying of inventory? Automating hundreds of hours of manual data gathering and reporting into a succinct/performant report. Why are we sending more inventory to warehouses that already have a significant % of our on hand quantity of that item? Building on Native AWS using CDK (Infrastructure as Code) and maintaining the with a CI/CD pipeline. Using AWS Sagemaker, can we predict how often a particular item will be within X miles of our customer at the time of order? What factors are driving our fulfillment optimization system to plan specific paths that will be used to fulfill customer demand for a particular fulfillment set?

  3. I work with some incredibly intelligent folks who I’ve learned a lot from.

  4. I have very few meetings each week, and the meetings I do have are generally quite interesting; I can actively contribute.

  5. I’ve never been told not to pursue something I’m interested in. I am constantly presented the opportunity to learn new things; whenever we are looking into a new project that I’m interested in working on (Not time sensitive) or I have an idea, I don’t get the responses: “Let’s leave that to the data science team”, “Lets let the Sr. BIE handle this one”. It’s more along the lines of “Yeah go for it, reach out to _ for help, and don’t delete any of our AWS resources!”. At this point in my career, I’m very grateful for this.


-I’d like to start writing on here more. I’m going to maintain these longer career posts
at their current cadence 1-2x per year, but I think I’ll start writing some
shorter technical pieces, maybe some about supply chain optimization problems,
and some personal thoughts I can look back on and see what I was thinking
as a 23 year old.

Relevant Links:

“Learn how to learn from those who offend you”