Writing
Case studies & field notes
What the work actually cost, what it saved, and how I know. Every number here traces to something you can check.
Two Shapes of Agent Orchestration: The Advisor and the Fan-OutTwo orchestration patterns cover nearly every multi-agent workflow I run: Executor + Advisor and Orchestrator + Workers. Here is where judgment belongs in each — and how to keep frontier models off grunt work.From Task Tickets to Platform Ownership: Four Years Inside One Marketplace BackendFounding engineer on one marketplace backend, owning the release path for four years — the arc from task tickets to platform ownership at Curbo. The evidence is the project's GitHub history.Insurance PDFs Are Hostile: An AI Extraction Pipeline in Three ActsAn eval-backed pipeline turned hostile insurance PDFs into structured data across an 82-state Step Functions workflow. It grew from model bake-off to API to serverless orchestration, grounded in a 179-PDF corpus.Retooling an Engineering Practice Around AI, One Repo at a TimeA small consultancy moved from internal bots to paid ML pipelines and production agents in three waves. Its repository timeline shows the shift — and the learning required to keep pace.Four AWS Certs, and Not One of Them Taught Me to Cut the BillFour AWS certifications opened doors; a production infrastructure audit taught me how to find what deserved to run. The lessons that moved money live in the audit trail, not the exams.How I Cut a Production AWS Bill by ~86% — From $25k to $3.5k a MonthAn ~86% reduction took production AWS from ~$25k to under $3.5k a month while the platform got faster. The gains came from database, compute, networking, dead-weight removal, and disciplined measurement.From Zip-and-Pray to Infrastructure as Code: Modernizing a Platform With No CI/CDAutomated infrastructure-as-code and CI/CD replaced manual zip uploads and an unknown cloud footprint. That foundation made the platform observable, repeatable, and ready for cost and reliability work.
Nothing under that tag yet.