Blog
Thinking on AI & software
Practical engineering articles from the people building these systems every day.
RAG vs. Fine-Tuning: Choosing the Right Approach for Your AI Product
Both techniques ground LLM outputs in your data, but they solve different problems. Here's how to decide which fits your use case.
What It Actually Takes to Run AI Agents in Production
Agent demos are easy. Reliable agents that touch real business systems are a different engineering problem entirely.
Five Cloud Cost Mistakes We See in Every Architecture Review
Most cloud spend problems aren't exotic — they're the same handful of mistakes compounding quietly over time.
Why Your Design System Should Ship Before Your First Feature
Teams that build a design system after the product exists spend years paying down inconsistency debt.
Multi-Tenant Architecture Decisions You Can't Undo Later
Some SaaS architecture decisions are cheap to change in month one and brutally expensive by year two.
Computer Vision Is Quietly Replacing Manual QA Lines
Defect detection models are now accurate enough to run unattended on production lines — here's what that rollout looks like.
Is Your Data Actually Ready for AI? A Practical Checklist
Before any model gets trained, most AI initiatives stall on the same data readiness gaps. Here's what to check first.
Want these insights in your inbox?
Subscribe to the Mabzone newsletter — new articles, twice a month.