ADINA WAGNER

Position: Research Associate (with a role spanning scientific coordination, research software engineering, and postdoctoral research)
Institute INM-7, Forschungszentrum Jülich
Academic Background: Clinical Psychology

Like many Research Software Engineers, Adina Wagner did not set out to become one. Trained as a clinical psychologist, she found her way into neuroinformatics and open-source research software through a passion for reproducible science and scientific collaboration.

Officially, Adina is a Research Associate. In practice, her role spans scientific coordination, research software engineering, and postdoctoral research. Although she identifies as a Research Software Engineer (RSE), the title she uses often depends on the context.

"I very much like the term RSE. However, the way I introduce myself depends on context. Often, I call myself a coordinator or Postdoc if those are more relevant to a task, event, or community at hand."

For Adina, being an RSE is defined less by her job title than by the work she does.

"Because I create and contribute to an open source ecosystem in science. Even if it's not the entirety of my job, it is the part I love the most."

The Journey

Research was always the part of psychology that fascinated Adina most.

"I realized early during my Psychology Bachelor that research was what was most interesting to me. I tried many flavours: classical psychological experiments in behavioral psychology, animal research on the cognition of corvids in the UK, clinical research in neurology and neuropsychology. But it was only when I discovered the field of neuroinformatics that I felt I found a true scientific home."

Her path into programming was gradual and largely self-directed.

"I had many unsuccessful attempts early on. During my psychology Bachelor I took extracurricular programming courses, but they didn't stick, until I self-taught myself R for data analysis. Those, I would say, are the first humble steps into coding, and I did them when I realized that this skill is necessary to continue in science."

The experience was very different from the software she had encountered during her studies.

"It was a stark contrast to what I was taught in university: mostly proprietary statistical tools like SPSS with shiny UIs and no proper way to codify your analyses."

During her master's thesis, she began learning Python.

"I started from scratch with Python for data analysis and visualization, and my first coding project: porting a MatLab toolbox into a Python package."

Before beginning her PhD, a research stay at Dartmouth College became another turning point.

"My advisor threw me head-first into the wonderful open source software ecosystem of the neuroimaging world."

In Practice

What Adina enjoys most is the opportunity to immerse herself in a problem from beginning to end.

"Whenever I find the time to dedicate a big chunk of time to really think a problem through from start to end, and have a piece of code eventually where I'm confident I understand the decisions that underlies it. Or a first-time contribution to a new project."

When she has a full day available for development, she has a routine.

"I typically try to tackle a bigger task from start to end. I work from home, which helps a lot with concentration. I start by either reading or writing an issue, and sharing some loose thoughts in a developer chat room for fast feedback. I commit and push changes into a work-in-progress pull request, so that others can read along if they want to. I really try to have at least a working prototype until I stop – then I can get meaningful feedback and it's easier for me and others to pick up the task later."

In reality, days like that are the exception.

"Mostly, I need to task switch between all kinds of work with only a fraction being actual coding."

The advice she offers to anyone developing research software is simple:

"First check what already exists, interact with that community and contribute or fix if necessary before re-implementing something new independently."

She also sees new challenges emerging across research software communities.

"Generative AI puts a strain on many open source software communities I interact with. It prevents new users or contributors from human-to-human interactions because people ask their agents to understand or contribute to packages—a valuable communication and onboarding path is slowly lost. It creates a human labour bottleneck when LLMs take seconds to create lengthy analyses and code contributions with hard to find bugs and often questionable implementations, but it needs a human maintainer with very little time to review those. And it creates social friction when there are opposing views on the use of LLMs in software ecosystems."

Current Work

Today, Adina contributes to the DataLad ecosystem, an open-source platform that integrates data, code, and metadata to support research data management and reproducibility.

Although she came to research software through neuroscience, her work now reaches researchers across a wide range of disciplines.

"It's a domain agnostic tool, so I interact with all kinds of disciplines: environmental science, material science, linguistics, even Archeology."

When asked what currently excites her most, her answer is brief but revealing.

"Interoperability."

She also values the opportunities her role at Forschungszentrum Jülich provides.

"FZJ gives me personally a lot of freedom, and a position where I can lobby and advocate for RSE work."

Community

For Adina, research software is about much more than writing code—it's about the people who build, maintain, and use it together.

"It is very useful to be connected and know 'your allies' in other German institutions. I also point any newcomer to these communities."

Her connection to JuRSE began even before the community formally existed.

"Before the community formally existed, Markus Diesmann from IAS-6 helped me connect with other people that cared for research software."

For her, strong software communities depend not only on good tools, but also on making the people behind them visible.

"The visibility of a community who codes is very valuable to me."

That perspective perhaps best captures both her journey into research software and what continues to motivate her today.

"I am a clinical psychologist, who turned to work with computers because they're not as complicated as people, but who realized very soon that the most important parts of software development are actually the people (users, developers, communities) and their social dynamics."

Quick Fire

Favourite programming language: Python

One tool you can't live without: Git

Best coding advice you've received: An old book on coding anti-patterns in C.