Michelle Lee

Data engineer in Los Angeles

I build the pipelines that move orders, returns and product data between Shopify, JD Edwards and Oracle at AG Jeans, and I run the CI platform those pipelines are scheduled on.

See my project Email me

Illustrated portrait of Michelle — Portrait by DJARN

What I do at work

AG Jeans designs its clothing, makes some of it in-house and sources most from vendors, then sells online, in its own stores and wholesale. Here’s how its data connects, and the parts I build and run.

Where we sell

agjeans.comagjeansoutlet.comRetail stores

Order integrations

Business systemthe company’s ERP and source of truth

Python automations

Where products and partner data live

Design system (PolyPM)Third-party appsVendor app, in progress

Sales and product data are also copied into a reporting database for the business to analyze.

Every connection runs on a schedule, on the platform I manage, watched by a monitoring app I built.

Keep every order flowingOrders, cancellations and returns from our two online stores and our retail stores reach the warehouse correctly.

Online and retail orders each run on their own integrations, and I own the ones for orders, cancellations and returns. When the company switched returns providers, I rebuilt the returns and exchanges process.

Shopify, JD Edwards, Oracle Integration Cloud

Keep product data consistentStyles, pricing and seasons match between the design team’s system and the business system.

These syncs run in both directions and are among our most complex automations. Sales and product data are also copied into a reporting database, set up so a rerun never creates duplicates.

Python, SQL, PolyPM, SQL Server, Oracle Autonomous Database

Connect the apps the business usesThird-party apps that support different sides of the business stay in sync with our systems.

Wholesale orders from department-store partners are imported from order files into the business system. In-store traffic data lands in the reporting database. New-customer forms load straight into the business system. I also build custom syncs for very specific data other teams rely on.

Python, REST APIs

Run the automation platformI manage the platform our scheduled jobs run on.

That covers who has access, where the code lives and shared pipeline templates. As part of a small team, I helped move 28 automations off an aging integration tool and rewrite them in Python.

GitLab, GitLab CI, Linux

Watch every scheduled jobOne app shows whether each job is turned on and when it last finished successfully.

Our jobs are scheduled across several systems, and GitLab only shows schedules one project at a time. I built a monitoring app that brings them all into one view.

GitLab, Oracle Cloud Infrastructure

Build new internal appsI’m building an app to support the company’s move to vendor-made clothing.

It’s built in Oracle APEX, the modern option we’re adopting for new business tools.

Oracle APEX, SQL

Train the teamI’m the go-to person for our code and platform, and I write the training.

That includes primers and hands-on exercises for Oracle Integration Cloud, the one part of our stack where AI coding tools can’t help.

Oracle Integration Cloud, AI-assisted coding

Outside work

A pipeline I built and run on my own time, held to the same standards as the ones I run at work.

Budget Pipeline

Running daily

Python · dbt · MotherDuck · Prefect · GitHub Actions · Cloudflare

Shopping is my most expensive hobby, and I’m Amazon’s best customer. Budgeting apps couldn’t tell me what I actually bought there, so I built my own: the cheapest option I could find that’s still secure, made for exactly how I spend. Every morning it pulls my bank and card activity, matches Amazon charges to the order history I export from Amazon, tests it all and updates a private dashboard with spending, goals and retirement projections.

I use it daily It replaced my budgeting app. I check it before I buy something, not just when I’m curious.
About $15 a year The only paid part is SimpleFIN, the service that connects to my banks ($15 a year). MotherDuck, GitHub Actions and Cloudflare all run on their free plans. That’s less than a typical budgeting app subscription.
Itemizes Amazon Every Amazon charge is split into the items in the order, with returns taken off, which budgeting apps don’t do.
Built around me Weekly paychecks, card due dates, savings goals without bank rules and retirement plans for different rents.

Read the case study Try the demo

Experience

Dec 2023 – nowData EngineerKoos Manufacturing (AG Jeans)
2022Web DeveloperCrescendo Interactive
2021B.S. Computer Information SystemsCal Poly Pomona

Before data work I built websites with React and JavaScript, which is where this site started in 2020.

Let’s talk

I’m interested in data engineering roles focused on pipelines and the platforms they run on. Email is the fastest way to reach me.