Engineering Notes

A public publication collection about AI-assisted engineering, workflow governance, and reliability.

AI-Assisted Engineering Reliability

This series asks how AI-assisted engineering can remain reviewable and trustworthy when generated output is not the same thing as workflow completion.

1. Prompt Is Not a Workflow Engine
2. Producing Work Is Not Completing Workflow
3. AI Can Produce Work, But What Counts as Evidence?
4. The Most Overlooked Problem in AI-Assisted Engineering
5. Completion Needs Evidence, Not Confidence
6. A Healthy Workflow Sometimes Says No
7. AI-Assisted Engineering Needs More Than Long Context

Read the series overview or browse all articles.

Research / Field Notes

Observations on emerging AI engineering practices, architectures, research, and reliability problems.

• AI Is Getting Better at Working. The Engineering Problem Is Changing Too.

Publication model

Each article remains readable as a repository-native Markdown file. The Pages layer provides a navigable presentation of that same content; it does not create a second copy of the article body.

• About this publication
• Research references
• Series documentation and article index
• Source repository on GitHub