At Micah Group, I built and ran ingestion for 9 prediction-market exchanges, sustaining 12.7M–17.9M price snapshots per day. I reduced missing index-input fields from 33% to 10% and archived 720M–1.1B rows without data loss.
I also wrote 5 SKILL.md workflows that cut new-exchange integration from about 2 weeks to 3 days. After profiling the archive pipeline, I changed compression and chunk sizes to cut transfer volume 17x.
Across electronic payments, banking, semiconductor inspection, and enterprise platforms, I’ve built backend services and APIs in Python, Node.js, and C#. I’ve also delivered front ends and automation, including a test framework that expanded coverage from 11% to 32% for an e-Beam inspection system.

