Better Code, Better Science¶
A free web book by Russell A. Poldrack (Stanford) on writing reproducible, high-quality scientific code: essential tools, testing and validation, project structure and data management, high-performance computing, sharing research objects, and (directly relevant here) AI-assisted coding workflows. Where Software Engineering Practices for Data Science is a quick checklist, this is the deeper read behind it.
Source: bettercode-book.org
- Free web edition, no signup, no paywall.
- Text is CC BY-NC-ND 4.0; code snippets are MIT-licensed and on GitHub (Better Code, Better Science org).
- The author is upfront that generative AI tools were used in researching and reviewing it, and that the example code was built through human-AI collaboration, worth reading as a model for how to disclose AI use in your own project write-ups.
Read the AI-assisted coding chapter alongside whichever track you're doing here: it's the same "review the agent's diff, don't just accept it" discipline this site's project briefs assume.