AI reading noteChip Huyen
AI Engineering
A guide to building applications around foundation models. It treats evaluation, prompt and context design, latency, cost, and deployment as connected engineering decisions.
AI Engineering
Key ideas
- Define an evaluation set before comparing models or prompts.
- Choose context and retrieval strategies from the task and evidence needs.
- Measure latency and cost alongside answer quality.
How I would apply it
Create a small golden set of real requests for every AI automation workflow.