Blog

Thinking about AI, architecture, and the future of building software.

AIAWSAgentsArchitectureAutomationEngineering LeadershipKnowledge GraphsPipelineProcessProductionProductivityPrompt EngineeringRAGSWEny
What a Knowledge Graph Is Actually For

A graph stores relationships, not passages. That single property decides what it is for, and what it should never be asked to do.
The AI-Forward SDLC: Put the Agents at Both Ends

The development lifecycle I built and run: zero-friction capture, agent-researched definition, parallel execution, and four tiers of validation.
The AI Software Factory: Team-Scale Output Without the Overhead

AI doesn't just help you code faster. It changes what a studio can realistically build and ship, and how fast.
Building an AI Pipeline with LLM Quality Gates

A deep dive into building a multi-stage AI generation pipeline on AWS: event-driven ingestion, LLM-as-judge evaluation, and migrating from SNS/SQS to Step Functions.
Building AI Assistants That Actually Work in Production

Lessons from shipping real AI assistant systems. Not chatbot demos, but production-grade tools that handle complex workflows.
Prompt Engineering for Production: Beyond the Playground

Production prompt engineering is about reliability, cost, and maintainability, not clever tricks.
Agentic Workflows: Making AI Do the Boring Work

How to design AI workflows that handle repetitive engineering tasks like triage, review, and deploy verification.