Technical writing on enterprise AI architecture, methodology, and deployment at scale.
Most enterprise-AI queries are relationship problems, not search problems. A two-tier, metadata-first knowledge graph answers most organizational questions with zero LLM cost, and inverts the cost structure of enterprise AI.Metadata-First Knowledge Graphs: Solving the Enterprise AI Context Problem
An idea is easy to describe and hard to believe. The two-day Rails + ngrok build of a restaurant-deals MVP that taught me to stop pitching and start shipping.Show, Don't Tell: Shipping a Restaurant-Deals MVP in Two Days
Corporate AI training fails because it's passive. Artifact-based learning, where every lesson produces real AI-generated work, is how you actually build workforce capability at scale.Why "Learn by Doing" Is the Only AI Training That Works
A production pattern for serverless AI pipelines: scheduled trigger, parallel extraction, structured LLM processing, and write-back, with the orchestration, error handling, and prompt engineering that make it production-grade.Event-Driven AI Pipelines: Automating Project Intelligence with Serverless Architecture
A Python tool that audits how every field of a Salesforce object is actually used, producing a CSV/Excel report with each field's last-used date and percentage of use, so admins can find and retire dead fields.Understanding the Field Usage of Any Object in Salesforce
When e-commerce demand grew 10x in two months and the data was trapped in a legacy back office, web scraping bridged the gap, feeding real-time capacity dashboards built with an agile, iterative approach.Web Scraping and Agile: How Web Scraping Helps You See the Big Picture