Rebuilding an Automotive AM Operation from Data Up — Yewon Hong
Preview
HL Mando Aftermarket North America │ Product & Technology Team │ Full-time, 2024–Present
TL;DR — Walked into a fragmented automotive product data operation and rebuilt it end-to-end: ERD-based data architecture, Python/VBA/UiPath automation, and a GenAI validation layer. Result: 40%+ fewer missing attributes, 10–15 hrs/week recovered from manual work, and a catalog that scales.
THE PROBLEM
The automotive aftermarket catalog operation ran on scattered spreadsheets, manual re-entry across Amazon VC / AAP / WHI, and no consistent data model. SKU–MSS–ASIN mappings were fragile. Amazon rejections were frequent. Promotion cycles required heroic effort to prepare. Nobody owned the full stack.
MY ROLE
I took ownership of the full lifecycle — not just one piece. Designed a relational ERD spanning SKU, MSS, ASIN, Fitment (ACES), and Attributes (PIES). Built automation scripts in Python, VBA, and UiPath to eliminate recurring manual tasks. Embedded a GenAI validation layer for attribute completeness and fitment accuracy. Coordinated between Amazon VSP, suppliers, sales, and engineering.