SERIES · Series 03
AI-paired engineering, in production.
Three parts on what AI-paired engineering looks like once the demo is over and the bill arrives.
THE PREMISE
Pair programming with an LLM is easy to do, and hard to do well at scale. This series covers the workflow shape, the eval suite that catches regressions before they reach the user, and the cost guardrails that make the math work in production.
- SD
Spec-driven development with LLMs
A spec the model can act on without rework. The structure, the anti-patterns, the actual diff.
9 min - ES
Eval suites for non-deterministic tools
How to test something that gives a different answer twice. The harness, the tolerances, the budget.
8 min - CG
Cost guardrails for AI workflows
A small accounting library that pays for itself in three weeks. Tokens, retries, and the budget alarm.
6 min
Done with the series?