Walking Backwards From Scientists: Lessons in Research Software Engineering

Most scientific software is built by researchers solving specific problems, resulting in powerful tools that remain inaccessible to the broader community due to poor usability and documentation. This means that there is a lot of untapped scientific potential locked away in inaccessible software. This talk presents a systematic methodology to unlock this potential by prioritizing both user and developer workflows alongside scientific objectives. Through examples from numerical relativity, climate science, and computational electromagnetics, I illustrate how investments in experience design--streamlined installation, clear documentation, intuitive interfaces--often yield substantial returns for scientific productivity, suggesting that user-centered design should be woven into research goals. The methodology provides actionable strategies for research software engineers seeking to balance scientific rigor with community accessibility, ultimately serving both scientific excellence and broader adoption.