At Amazon and Convoy I specialized in performance: taking systems that already worked and finding out what they were quietly wasting. Nothing was "broken." Everything was expensive.

Audited and optimized legacy SQL queries: cut database utilization by 50% and made our biggest hourly job 500× faster — from 2 hours down to 15 seconds

Optimized big-data pipelines to cut compute and storage costs by 75%, saving $600,000 every month

Rebuilt a bulk storage update path: 1 million records updated in 1.5 minutes, down from 60

Built a Python tool that auto-tuned Spark job parameters for the best speed-to-cost ratio — turning optimization from a one-off effort into an automated system

Trained the team on SQL optimization, so the discipline outlived my involvement

An audit's bet is always the same: a system that runs is not a system that runs well, and the waste stays invisible until someone measures it. That was true of pipelines at Amazon, and it's true of a Shopify store — slow pages, a leaky checkout, ad spend that quietly stopped working. Finding money in systems people assumed were fine is the job. The surfaces changed; the bet hasn't.